fix: felt to IntegerRep overflow should panic

BREAKING CHANGE: compiled model serialization is not backwards compatible

.cargo/config

@@ -1,4 +0,0 @@
-[target.wasm32-unknown-unknown]
-runner = 'wasm-bindgen-test-runner'
-rustflags = ["-C", "target-feature=+atomics,+bulk-memory,+mutable-globals","-C",
- "link-arg=--max-memory=4294967296"]

.cargo/config.toml

@@ -0,0 +1,17 @@
+[target.wasm32-unknown-unknown]
+runner = 'wasm-bindgen-test-runner'
+rustflags = ["-C", "target-feature=+atomics,+bulk-memory,+mutable-globals","-C",
+ "link-arg=--max-memory=4294967296"]
+
+
+[target.x86_64-apple-darwin]
+rustflags = [
+  "-C", "link-arg=-undefined",
+  "-C", "link-arg=dynamic_lookup",
+]
+
+[target.aarch64-apple-darwin]
+rustflags = [
+  "-C", "link-arg=-undefined",
+  "-C", "link-arg=dynamic_lookup",
+]

.github/workflows/benchmarks.yml

@@ -6,23 +6,16 @@ on:
         description: "Test scenario tags"
 
 jobs:
-  bench_elgamal:
-    runs-on: self-hosted
-    steps:
-      - uses: actions/checkout@v4
-      - uses: actions-rs/toolchain@v1
-        with:
-          toolchain: nightly-2023-06-27
-          override: true
-          components: rustfmt, clippy
-      - name: Bench elgamal
-        run: cargo bench --verbose --bench elgamal
 
   bench_poseidon:
+    permissions:
+      contents: read
     runs-on: self-hosted
     steps:
-      - uses: actions/checkout@v4
-      - uses: actions-rs/toolchain@v1
+      - uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683  #v4.2.2
+        with:
+          persist-credentials: false
+      - uses: actions-rs/toolchain@b2417cde72dcf67f306c0ae8e0828a81bf0b189f  #v1.0.6
         with:
           toolchain: nightly-2023-06-27
           override: true
@@ -31,11 +24,15 @@ jobs:
         run: cargo bench --verbose --bench poseidon
 
   bench_einsum_accum_matmul:
+    permissions:
+      contents: read
     runs-on: self-hosted
     needs: [bench_poseidon]
     steps:
-      - uses: actions/checkout@v4
-      - uses: actions-rs/toolchain@v1
+      - uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683  #v4.2.2
+        with:
+          persist-credentials: false
+      - uses: actions-rs/toolchain@b2417cde72dcf67f306c0ae8e0828a81bf0b189f  #v1.0.6
         with:
           toolchain: nightly-2023-06-27
           override: true
@@ -44,11 +41,15 @@ jobs:
         run: cargo bench --verbose --bench accum_einsum_matmul
 
   bench_accum_matmul_relu:
+    permissions:
+      contents: read
     runs-on: self-hosted
     needs: [bench_poseidon]
     steps:
-      - uses: actions/checkout@v4
-      - uses: actions-rs/toolchain@v1
+      - uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683  #v4.2.2
+        with:
+          persist-credentials: false
+      - uses: actions-rs/toolchain@b2417cde72dcf67f306c0ae8e0828a81bf0b189f  #v1.0.6
         with:
           toolchain: nightly-2023-06-27
           override: true
@@ -57,11 +58,15 @@ jobs:
         run: cargo bench --verbose --bench accum_matmul_relu
 
   bench_accum_matmul_relu_overflow:
+    permissions:
+      contents: read
     runs-on: self-hosted
     needs: [bench_poseidon]
     steps:
-      - uses: actions/checkout@v4
-      - uses: actions-rs/toolchain@v1
+      - uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683  #v4.2.2
+        with:
+          persist-credentials: false
+      - uses: actions-rs/toolchain@b2417cde72dcf67f306c0ae8e0828a81bf0b189f  #v1.0.6
         with:
           toolchain: nightly-2023-06-27
           override: true
@@ -70,11 +75,15 @@ jobs:
         run: cargo bench --verbose --bench accum_matmul_relu_overflow
 
   bench_relu:
+    permissions:
+      contents: read
     runs-on: self-hosted
     needs: [bench_poseidon]
     steps:
-      - uses: actions/checkout@v4
-      - uses: actions-rs/toolchain@v1
+      - uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683  #v4.2.2
+        with:
+          persist-credentials: false
+      - uses: actions-rs/toolchain@b2417cde72dcf67f306c0ae8e0828a81bf0b189f  #v1.0.6
         with:
           toolchain: nightly-2023-06-27
           override: true
@@ -83,11 +92,15 @@ jobs:
         run: cargo bench --verbose --bench relu
 
   bench_accum_dot:
+    permissions:
+      contents: read
     runs-on: self-hosted
     needs: [bench_poseidon]
     steps:
-      - uses: actions/checkout@v4
-      - uses: actions-rs/toolchain@v1
+      - uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683  #v4.2.2
+        with:
+          persist-credentials: false
+      - uses: actions-rs/toolchain@b2417cde72dcf67f306c0ae8e0828a81bf0b189f  #v1.0.6
         with:
           toolchain: nightly-2023-06-27
           override: true
@@ -96,11 +109,15 @@ jobs:
         run: cargo bench --verbose --bench accum_dot
 
   bench_accum_conv:
+    permissions:
+      contents: read
     runs-on: self-hosted
     needs: [bench_poseidon]
     steps:
-      - uses: actions/checkout@v4
-      - uses: actions-rs/toolchain@v1
+      - uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683  #v4.2.2
+        with:
+          persist-credentials: false
+      - uses: actions-rs/toolchain@b2417cde72dcf67f306c0ae8e0828a81bf0b189f  #v1.0.6
         with:
           toolchain: nightly-2023-06-27
           override: true
@@ -109,11 +126,15 @@ jobs:
         run: cargo bench --verbose --bench accum_conv
 
   bench_accum_sumpool:
+    permissions:
+      contents: read
     runs-on: self-hosted
     needs: [bench_poseidon]
     steps:
-      - uses: actions/checkout@v4
-      - uses: actions-rs/toolchain@v1
+      - uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683  #v4.2.2
+        with:
+          persist-credentials: false
+      - uses: actions-rs/toolchain@b2417cde72dcf67f306c0ae8e0828a81bf0b189f  #v1.0.6
         with:
           toolchain: nightly-2023-06-27
           override: true
@@ -122,11 +143,15 @@ jobs:
         run: cargo bench --verbose --bench accum_sumpool
 
   bench_pairwise_add:
+    permissions:
+      contents: read
     runs-on: self-hosted
     needs: [bench_poseidon]
     steps:
-      - uses: actions/checkout@v4
-      - uses: actions-rs/toolchain@v1
+      - uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683  #v4.2.2
+        with:
+          persist-credentials: false
+      - uses: actions-rs/toolchain@b2417cde72dcf67f306c0ae8e0828a81bf0b189f  #v1.0.6
         with:
           toolchain: nightly-2023-06-27
           override: true
@@ -135,11 +160,15 @@ jobs:
         run: cargo bench --verbose --bench pairwise_add
 
   bench_accum_sum:
+    permissions:
+      contents: read
     runs-on: self-hosted
     needs: [bench_poseidon]
     steps:
-      - uses: actions/checkout@v4
-      - uses: actions-rs/toolchain@v1
+      - uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683  #v4.2.2
+        with:
+          persist-credentials: false
+      - uses: actions-rs/toolchain@b2417cde72dcf67f306c0ae8e0828a81bf0b189f  #v1.0.6
         with:
           toolchain: nightly-2023-06-27
           override: true
@@ -148,11 +177,15 @@ jobs:
         run: cargo bench --verbose --bench accum_sum
 
   bench_pairwise_pow:
+    permissions:
+      contents: read
     runs-on: self-hosted
     needs: [bench_poseidon]
     steps:
-      - uses: actions/checkout@v4
-      - uses: actions-rs/toolchain@v1
+      - uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683  #v4.2.2
+        with:
+          persist-credentials: false
+      - uses: actions-rs/toolchain@b2417cde72dcf67f306c0ae8e0828a81bf0b189f  #v1.0.6
         with:
           toolchain: nightly-2023-06-27
           override: true

.github/workflows/engine.yml

@@ -15,22 +15,32 @@ defaults:
     working-directory: .
 jobs:
   publish-wasm-bindings:
+    permissions:
+      contents: read
+      packages: write
     name: publish-wasm-bindings
+    env:
+      RELEASE_TAG: ${{ github.ref_name }}
     runs-on: ubuntu-latest
     if: startsWith(github.ref, 'refs/tags/')
     steps:
-      - uses: actions/checkout@v4
-      - uses: actions-rs/toolchain@v1
+      - uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683  #v4.2.2
         with:
-          toolchain: nightly-2024-02-06
+          persist-credentials: false
+      - uses: actions-rs/toolchain@b2417cde72dcf67f306c0ae8e0828a81bf0b189f  #v1.0.6
+        with:
+          toolchain: nightly-2024-07-18
           override: true
           components: rustfmt, clippy
-      - uses: jetli/wasm-pack-action@v0.4.0
+      - uses: jetli/wasm-pack-action@0d096b08b4e5a7de8c28de67e11e945404e9eefa  #v0.4.0
+        with:
+          # Pin to version 0.12.1
+          version: 'v0.12.1'
       - name: Add wasm32-unknown-unknown target
         run: rustup target add wasm32-unknown-unknown
 
       - name: Add rust-src
-        run: rustup component add rust-src --toolchain nightly-2024-02-06-x86_64-unknown-linux-gnu
+        run: rustup component add rust-src --toolchain nightly-2024-07-18-x86_64-unknown-linux-gnu
       - name: Install binaryen
         run: |
           set -e
@@ -39,41 +49,41 @@ jobs:
           wasm-opt --version
       - name: Build wasm files for both web and nodejs compilation targets
         run: |
-          wasm-pack build --release --target nodejs --out-dir ./pkg/nodejs . -- -Z build-std="panic_abort,std"
-          wasm-pack build --release --target web --out-dir ./pkg/web . -- -Z build-std="panic_abort,std" --features web
+            wasm-pack build --release --target nodejs --out-dir ./pkg/nodejs . -- -Z build-std="panic_abort,std"
+            wasm-pack build --release --target web --out-dir ./pkg/web . -- -Z build-std="panic_abort,std" --features web
       - name: Create package.json in pkg folder
         shell: bash
-        env:
-          RELEASE_TAG: ${{ github.ref_name }}
         run: |
-          echo '{
-            "name": "@ezkljs/engine",
-            "version": "${{ github.ref_name }}",
-            "dependencies": {
-              "@types/json-bigint": "^1.0.1",
-              "json-bigint": "^1.0.0"
-            },
-            "files": [
-              "nodejs/ezkl_bg.wasm",
-              "nodejs/ezkl.js",
-              "nodejs/ezkl.d.ts",
-              "nodejs/package.json",
-              "nodejs/utils.js",
-              "web/ezkl_bg.wasm",
-              "web/ezkl.js",
-              "web/ezkl.d.ts",
-              "web/snippets/**/*",
-              "web/package.json",
-              "web/utils.js",
-              "ezkl.d.ts"
-            ],
-            "main": "nodejs/ezkl.js",
-            "module": "web/ezkl.js",
-            "types": "nodejs/ezkl.d.ts",
-            "sideEffects": [
-              "web/snippets/*"
-            ]
-          }' > pkg/package.json
+            cat > pkg/package.json << EOF
+            {
+              "name": "@ezkljs/engine",
+              "version": "$RELEASE_TAG",
+              "dependencies": {
+                "@types/json-bigint": "^1.0.1",
+                "json-bigint": "^1.0.0"
+              },
+              "files": [
+                "nodejs/ezkl_bg.wasm",
+                "nodejs/ezkl.js",
+                "nodejs/ezkl.d.ts",
+                "nodejs/package.json",
+                "nodejs/utils.js",
+                "web/ezkl_bg.wasm",
+                "web/ezkl.js",
+                "web/ezkl.d.ts",
+                "web/snippets/**/*",
+                "web/package.json",
+                "web/utils.js",
+                "ezkl.d.ts"
+              ],
+              "main": "nodejs/ezkl.js",
+              "module": "web/ezkl.js",
+              "types": "nodejs/ezkl.d.ts",
+              "sideEffects": [
+                "web/snippets/*"
+              ]
+            }
+            EOF
 
       - name: Replace memory definition in nodejs
         run: |
@@ -166,7 +176,7 @@ jobs:
           curl -s "https://raw.githubusercontent.com/zkonduit/ezkljs-engine/main/README.md" > ./pkg/README.md
 
       - name: Set up Node.js
-        uses: actions/setup-node@v3
+        uses: actions/setup-node@1a4442cacd436585916779262731d5b162bc6ec7  #v3.8.2
         with:
           node-version: "18.12.1"
           registry-url: "https://registry.npmjs.org"
@@ -181,21 +191,26 @@ jobs:
 
 
   in-browser-evm-ver-publish:
+    permissions:
+      contents: read
+      packages: write
     name: publish-in-browser-evm-verifier-package
     needs: [publish-wasm-bindings]
     runs-on: ubuntu-latest
+    env:
+        RELEASE_TAG: ${{ github.ref_name }}
     if: startsWith(github.ref, 'refs/tags/')
     steps:
-      - uses: actions/checkout@v4
+      - uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683  #v4.2.2
+        with:
+          persist-credentials: false
       - name: Update version in package.json
         shell: bash
-        env:
-          RELEASE_TAG: ${{ github.ref_name }}
         run: |
-          sed -i "s|\"version\": \".*\"|\"version\": \"${{ github.ref_name }}\"|" in-browser-evm-verifier/package.json
+          sed -i "s|\"version\": \".*\"|\"version\": \"$RELEASE_TAG\"|" in-browser-evm-verifier/package.json
       - name: Prepare tag and fetch package integrity
         run: |
-          CLEANED_TAG=${{ github.ref_name }} # Get the tag from ref_name
+          CLEANED_TAG=${RELEASE_TAG} # Get the tag from ref_name
           CLEANED_TAG="${CLEANED_TAG#v}" # Remove leading 'v'
           echo "CLEANED_TAG=${CLEANED_TAG}" >> $GITHUB_ENV # Set it as an environment variable for later steps
           ENGINE_INTEGRITY=$(npm view @ezkljs/engine@$CLEANED_TAG dist.integrity)
@@ -215,13 +230,13 @@ jobs:
           NR==30{$0="    specifier: \"" tag "\""}
           NR==31{$0="    version: \"" tag "\""}
           NR==400{$0="  /@ezkljs/engine@" tag ":"}
-          NR==401{$0="    resolution: {integrity: \"" integrity "\"}"} 1' in-browser-evm-verifier/pnpm-lock.yaml > temp.yaml && mv temp.yaml in-browser-evm-verifier/pnpm-lock.yaml  
+          NR==401{$0="    resolution: {integrity: \"" integrity "\"}"} 1' in-browser-evm-verifier/pnpm-lock.yaml > temp.yaml && mv temp.yaml in-browser-evm-verifier/pnpm-lock.yaml
       - name: Use pnpm 8
-        uses: pnpm/action-setup@v2
+        uses: pnpm/action-setup@eae0cfeb286e66ffb5155f1a79b90583a127a68b  #v2.4.1
         with:
           version: 8
       - name: Set up Node.js
-        uses: actions/setup-node@v3
+        uses: actions/setup-node@1a4442cacd436585916779262731d5b162bc6ec7  #v3.8.2
         with:
           node-version: "18.12.1"
           registry-url: "https://registry.npmjs.org"
@@ -232,4 +247,4 @@ jobs:
           pnpm run build
           pnpm publish --no-git-checks
         env:
-          NODE_AUTH_TOKEN: ${{ secrets.NPM_TOKEN }}  
+          NODE_AUTH_TOKEN: ${{ secrets.NPM_TOKEN }}

.github/workflows/large-tests.yml

@@ -6,12 +6,16 @@ on:
         description: "Test scenario tags"
 jobs:
   large-tests:
+    permissions:
+      contents: read
     runs-on: kaiju
     steps:
-      - uses: actions/checkout@v4
-      - uses: actions-rs/toolchain@v1
+      - uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683  #v4.2.2
         with:
-          toolchain: nightly-2024-02-06
+          persist-credentials: false
+      - uses: actions-rs/toolchain@b2417cde72dcf67f306c0ae8e0828a81bf0b189f  #v1.0.6
+        with:
+          toolchain: nightly-2024-07-18
           override: true
           components: rustfmt, clippy
       - name: nanoGPT Mock

.github/workflows/pypi-gpu.yml

@@ -18,38 +18,46 @@ defaults:
 jobs:
 
   linux:
+    permissions:
+      contents: read
+      packages: write
     runs-on: GPU
     strategy:
       matrix:
         target: [x86_64]
+    env:
+      RELEASE_TAG: ${{ github.ref_name }}
     steps:
-      - uses: actions/checkout@v4
-      - uses: actions/setup-python@v4
+      - uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683  #v4.2.2
+        with:
+          persist-credentials: false
+      - uses: actions/setup-python@b64ffcaf5b410884ad320a9cfac8866006a109aa  #v4.8.0
         with:
           python-version: 3.12
           architecture: x64
 
-      - name: Set pyproject.toml version to match github tag
+      - name: Set pyproject.toml version to match github tag and rename ezkl to ezkl-gpu
         shell: bash
         run: |
             mv pyproject.toml pyproject.toml.orig
-            sed "s/ezkl/ezkl-gpu/" pyproject.toml.orig >pyproject.toml
+            sed "s/ezkl/ezkl-gpu/" pyproject.toml.orig > pyproject.toml.tmp
+            sed "s/0\\.0\\.0/${RELEASE_TAG//v}/" pyproject.toml.tmp > pyproject.toml
 
-      - uses: actions-rs/toolchain@v1
+      - uses: actions-rs/toolchain@b2417cde72dcf67f306c0ae8e0828a81bf0b189f  #v1.0.6
         with:
           toolchain: nightly-2023-06-27
           override: true
           components: rustfmt, clippy
 
-      - name: Set Cargo.toml version to match github tag
+      - name: Set Cargo.toml version to match github tag and rename ezkl to ezkl-gpu
         shell: bash
-        env:
-          RELEASE_TAG: ${{ github.ref_name }}
+        # the ezkl substitution here looks for the first instance of name = "ezkl" and changes it to "ezkl-gpu"
         run: |
           mv Cargo.toml Cargo.toml.orig
-          sed "s/0\\.0\\.0/${RELEASE_TAG//v}/" Cargo.toml.orig >Cargo.toml
+          sed "0,/name = \"ezkl\"/s/name = \"ezkl\"/name = \"ezkl-gpu\"/" Cargo.toml.orig > Cargo.toml.tmp
+          sed "s/0\\.0\\.0/${RELEASE_TAG//v}/" Cargo.toml.tmp > Cargo.toml
           mv Cargo.lock Cargo.lock.orig
-          sed "s/0\\.0\\.0/${RELEASE_TAG//v}/" Cargo.lock.orig >Cargo.lock
+          sed "s/0\\.0\\.0/${RELEASE_TAG//v}/" Cargo.lock.orig > Cargo.lock
 
       - name: Install required libraries
         shell: bash
@@ -57,7 +65,7 @@ jobs:
           sudo apt-get update && sudo apt-get install -y openssl pkg-config libssl-dev
 
       - name: Build wheels
-        uses: PyO3/maturin-action@v1
+        uses: PyO3/maturin-action@5f8a1b3b0aad13193f46c9131f9b9e663def8ce5  #v1.46.0
         with:
           target: ${{ matrix.target }}
           manylinux: auto
@@ -70,7 +78,7 @@ jobs:
           pip install ezkl-gpu --no-index --find-links dist --force-reinstall
 
       - name: Upload wheels
-        uses: actions/upload-artifact@v3
+        uses: actions/upload-artifact@65c4c4a1ddee5b72f698fdd19549f0f0fb45cf08  #v4.6.0
         with:
           name: wheels
           path: dist
@@ -86,7 +94,7 @@ jobs:
     # needs: [ macos, windows, linux, linux-cross, musllinux, musllinux-cross ]
     needs: [linux]
     steps:
-      - uses: actions/download-artifact@v3
+      - uses: actions/download-artifact@fa0a91b85d4f404e444e00e005971372dc801d16  #v4.1.8
         with:
           name: wheels
       - name: List Files
@@ -98,14 +106,14 @@ jobs:
       # publishes to PyPI
       - name: Publish package distributions to PyPI
         continue-on-error: true
-        uses: pypa/gh-action-pypi-publish@release/v1
+        uses: pypa/gh-action-pypi-publish@76f52bc884231f62b9a034ebfe128415bbaabdfc  #v1.12.4
         with:
-          packages-dir: ./
+          packages-dir: ./wheels
 
       # publishes to TestPyPI
       - name: Publish package distribution to TestPyPI
         continue-on-error: true
-        uses: pypa/gh-action-pypi-publish@release/v1
+        uses: pypa/gh-action-pypi-publish@76f52bc884231f62b9a034ebfe128415bbaabdfc  #v1.12.4
         with:
           repository-url: https://test.pypi.org/legacy/
-          packages-dir: ./
+          packages-dir: ./wheels

.github/workflows/pypi.yml

@@ -16,36 +16,53 @@ defaults:
 
 jobs:
   macos:
+    permissions:
+      contents: read
     runs-on: macos-latest
     if: startsWith(github.ref, 'refs/tags/')
     strategy:
       matrix:
         target: [x86_64, universal2-apple-darwin]
+    env:
+        RELEASE_TAG: ${{ github.ref_name }}
     steps:
-      - uses: actions/checkout@v4
-      - uses: actions/setup-python@v4
+      - uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683  #v4.2.2
+        with:
+          persist-credentials: false
+      - uses: actions/setup-python@b64ffcaf5b410884ad320a9cfac8866006a109aa  #v4.8.0
         with:
           python-version: 3.12
           architecture: x64
 
+      - name: Set pyproject.toml version to match github tag
+        shell: bash
+        run: |
+          mv pyproject.toml pyproject.toml.orig
+          sed "s/0\\.0\\.0/${RELEASE_TAG//v}/" pyproject.toml.orig >pyproject.toml
+
       - name: Set Cargo.toml version to match github tag
         shell: bash
-        env:
-          RELEASE_TAG: ${{ github.ref_name }}
         run: |
           mv Cargo.toml Cargo.toml.orig
           sed "s/0\\.0\\.0/${RELEASE_TAG//v}/" Cargo.toml.orig >Cargo.toml
           mv Cargo.lock Cargo.lock.orig
           sed "s/0\\.0\\.0/${RELEASE_TAG//v}/" Cargo.lock.orig >Cargo.lock
 
-      - uses: actions-rs/toolchain@v1
+      - uses: actions-rs/toolchain@b2417cde72dcf67f306c0ae8e0828a81bf0b189f  #v1.0.6
         with:
-          toolchain: nightly-2024-02-06
+          toolchain: nightly-2024-07-18
           override: true
           components: rustfmt, clippy
 
       - name: Build wheels
-        uses: PyO3/maturin-action@v1
+        if: matrix.target == 'universal2-apple-darwin'
+        uses: PyO3/maturin-action@5f8a1b3b0aad13193f46c9131f9b9e663def8ce5  #v1.46.0
+        with:
+          target: ${{ matrix.target }}
+          args: --release --out dist --features python-bindings
+      - name: Build wheels
+        if: matrix.target == 'x86_64'
+        uses: PyO3/maturin-action@5f8a1b3b0aad13193f46c9131f9b9e663def8ce5  #v1.46.0
         with:
           target: ${{ matrix.target }}
           args: --release --out dist --features python-bindings
@@ -56,24 +73,36 @@ jobs:
           python -c "import ezkl"
 
       - name: Upload wheels
-        uses: actions/upload-artifact@v3
+        uses: actions/upload-artifact@65c4c4a1ddee5b72f698fdd19549f0f0fb45cf08  #v4.6.0
         with:
-          name: wheels
+          name: dist-macos-${{ matrix.target }}
           path: dist
 
   windows:
+    permissions:
+      contents: read
     runs-on: windows-latest
     if: startsWith(github.ref, 'refs/tags/')
     strategy:
       matrix:
         target: [x64, x86]
     steps:
-      - uses: actions/checkout@v4
-      - uses: actions/setup-python@v4
+      - uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683  #v4.2.2
+        with:
+          persist-credentials: false
+      - uses: actions/setup-python@b64ffcaf5b410884ad320a9cfac8866006a109aa  #v4.8.0
         with:
           python-version: 3.12
           architecture: ${{ matrix.target }}
 
+      - name: Set pyproject.toml version to match github tag
+        shell: bash
+        env:
+          RELEASE_TAG: ${{ github.ref_name }}
+        run: |
+          mv pyproject.toml pyproject.toml.orig
+          sed "s/0\\.0\\.0/${RELEASE_TAG//v}/" pyproject.toml.orig >pyproject.toml
+
       - name: Set Cargo.toml version to match github tag
         shell: bash
         env:
@@ -84,14 +113,14 @@ jobs:
           mv Cargo.lock Cargo.lock.orig
           sed "s/0\\.0\\.0/${RELEASE_TAG//v}/" Cargo.lock.orig >Cargo.lock
 
-      - uses: actions-rs/toolchain@v1
+      - uses: actions-rs/toolchain@b2417cde72dcf67f306c0ae8e0828a81bf0b189f  #v1.0.6
         with:
-          toolchain: nightly-2024-02-06
+          toolchain: nightly-2024-07-18
           override: true
           components: rustfmt, clippy
 
       - name: Build wheels
-        uses: PyO3/maturin-action@v1
+        uses: PyO3/maturin-action@5f8a1b3b0aad13193f46c9131f9b9e663def8ce5  #v1.46.0
         with:
           target: ${{ matrix.target }}
           args: --release --out dist --features python-bindings
@@ -101,24 +130,36 @@ jobs:
           python -c "import ezkl"
 
       - name: Upload wheels
-        uses: actions/upload-artifact@v3
+        uses: actions/upload-artifact@65c4c4a1ddee5b72f698fdd19549f0f0fb45cf08  #v4.6.0 #v4.6.0
         with:
-          name: wheels
+          name: dist-windows-${{ matrix.target }}
           path: dist
 
   linux:
+    permissions:
+      contents: read
     runs-on: ubuntu-latest
     if: startsWith(github.ref, 'refs/tags/')
     strategy:
       matrix:
         target: [x86_64]
     steps:
-      - uses: actions/checkout@v4
-      - uses: actions/setup-python@v4
+      - uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683  #v4.2.2
+        with:
+          persist-credentials: false
+      - uses: actions/setup-python@b64ffcaf5b410884ad320a9cfac8866006a109aa  #v4.8.0
         with:
           python-version: 3.12
           architecture: x64
 
+      - name: Set pyproject.toml version to match github tag
+        shell: bash
+        env:
+          RELEASE_TAG: ${{ github.ref_name }}
+        run: |
+          mv pyproject.toml pyproject.toml.orig
+          sed "s/0\\.0\\.0/${RELEASE_TAG//v}/" pyproject.toml.orig >pyproject.toml
+
       - name: Set Cargo.toml version to match github tag
         shell: bash
         env:
@@ -129,14 +170,13 @@ jobs:
           mv Cargo.lock Cargo.lock.orig
           sed "s/0\\.0\\.0/${RELEASE_TAG//v}/" Cargo.lock.orig >Cargo.lock
 
-
       - name: Install required libraries
         shell: bash
         run: |
           sudo apt-get update && sudo apt-get install -y openssl pkg-config libssl-dev
 
       - name: Build wheels
-        uses: PyO3/maturin-action@v1
+        uses: PyO3/maturin-action@5f8a1b3b0aad13193f46c9131f9b9e663def8ce5  #v1.46.0
         with:
           target: ${{ matrix.target }}
           manylinux: auto
@@ -163,63 +203,14 @@ jobs:
           python -c "import ezkl"
 
       - name: Upload wheels
-        uses: actions/upload-artifact@v3
+        uses: actions/upload-artifact@65c4c4a1ddee5b72f698fdd19549f0f0fb45cf08  #v4.6.0
         with:
-          name: wheels
+          name: dist-linux-${{ matrix.target }}
           path: dist
 
-  # TODO: There's a problem with the maturin-action toolchain for arm arch leading to failed builds
-  # linux-cross:
-  #   runs-on: ubuntu-latest
-  #   strategy:
-  #     matrix:
-  #       target: [aarch64, armv7]
-  #   steps:
-  #   - uses: actions/checkout@v4
-  #   - uses: actions/setup-python@v4
-  #     with:
-  #       python-version: 3.12
-
-  #   - name: Install cross-compilation tools for aarch64
-  #     if: matrix.target == 'aarch64'
-  #     run: |
-  #       sudo apt-get update
-  #       sudo apt-get install -y gcc make gcc-aarch64-linux-gnu binutils-aarch64-linux-gnu libc6-dev-arm64-cross libusb-1.0-0-dev libatomic1-arm64-cross
-
-  #   - name: Install cross-compilation tools for armv7
-  #     if: matrix.target == 'armv7'
-  #     run: |
-  #       sudo apt-get update
-  #       sudo apt-get install -y gcc make gcc-arm-linux-gnueabihf binutils-arm-linux-gnueabihf libc6-dev-armhf-cross libusb-1.0-0-dev libatomic1-armhf-cross
-
-  #   - name: Build wheels
-  #     uses: PyO3/maturin-action@v1
-  #     with:
-  #       target: ${{ matrix.target }}
-  #       manylinux: auto
-  #       args: --release --out dist --features python-bindings
-
-  #   - uses: uraimo/run-on-arch-action@v2.5.0
-  #     name: Install built wheel
-  #     with:
-  #       arch: ${{ matrix.target }}
-  #       distro: ubuntu20.04
-  #       githubToken: ${{ github.token }}
-  #       install: |
-  #         apt-get update
-  #         apt-get install -y --no-install-recommends python3 python3-pip
-  #         pip3 install -U pip
-  #       run: |
-  #         pip3 install ezkl --no-index --find-links dist/ --force-reinstall
-  #         python3 -c "import ezkl"
-
-  #   - name: Upload wheels
-  #     uses: actions/upload-artifact@v3
-  #     with:
-  #       name: wheels
-  #       path: dist
-
   musllinux:
+    permissions:
+      contents: read
     runs-on: ubuntu-latest
     if: startsWith(github.ref, 'refs/tags/')
     strategy:
@@ -227,12 +218,22 @@ jobs:
         target:
           - x86_64-unknown-linux-musl
     steps:
-      - uses: actions/checkout@v4
-      - uses: actions/setup-python@v4
+      - uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683  #v4.2.2
+        with:
+          persist-credentials: false
+      - uses: actions/setup-python@b64ffcaf5b410884ad320a9cfac8866006a109aa  #v4.8.0
         with:
           python-version: 3.12
           architecture: x64
 
+      - name: Set pyproject.toml version to match github tag
+        shell: bash
+        env:
+          RELEASE_TAG: ${{ github.ref_name }}
+        run: |
+          mv pyproject.toml pyproject.toml.orig
+          sed "s/0\\.0\\.0/${RELEASE_TAG//v}/" pyproject.toml.orig >pyproject.toml
+
       - name: Set Cargo.toml version to match github tag
         shell: bash
         env:
@@ -249,7 +250,7 @@ jobs:
           sudo apt-get update && sudo apt-get install -y pkg-config libssl-dev
 
       - name: Build wheels
-        uses: PyO3/maturin-action@v1
+        uses: PyO3/maturin-action@5f8a1b3b0aad13193f46c9131f9b9e663def8ce5  #v1.46.0
         with:
           target: ${{ matrix.target }}
           manylinux: musllinux_1_2
@@ -270,12 +271,14 @@ jobs:
             python3 -c "import ezkl"
 
       - name: Upload wheels
-        uses: actions/upload-artifact@v3
+        uses: actions/upload-artifact@65c4c4a1ddee5b72f698fdd19549f0f0fb45cf08  #v4.6.0
         with:
-          name: wheels
+          name: dist-musllinux-${{ matrix.target }}
           path: dist
 
   musllinux-cross:
+    permissions:
+      contents: read
     runs-on: ubuntu-latest
     if: startsWith(github.ref, 'refs/tags/')
     strategy:
@@ -283,14 +286,22 @@ jobs:
         platform:
           - target: aarch64-unknown-linux-musl
             arch: aarch64
-          - target: armv7-unknown-linux-musleabihf
-            arch: armv7
     steps:
-      - uses: actions/checkout@v4
-      - uses: actions/setup-python@v4
+      - uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683  #v4.2.2
+        with:
+          persist-credentials: false
+      - uses: actions/setup-python@b64ffcaf5b410884ad320a9cfac8866006a109aa  #v4.8.0
         with:
           python-version: 3.12
 
+      - name: Set pyproject.toml version to match github tag
+        shell: bash
+        env:
+          RELEASE_TAG: ${{ github.ref_name }}
+        run: |
+          mv pyproject.toml pyproject.toml.orig
+          sed "s/0\\.0\\.0/${RELEASE_TAG//v}/" pyproject.toml.orig >pyproject.toml
+
       - name: Set Cargo.toml version to match github tag
         shell: bash
         env:
@@ -302,13 +313,13 @@ jobs:
           sed "s/0\\.0\\.0/${RELEASE_TAG//v}/" Cargo.lock.orig >Cargo.lock
 
       - name: Build wheels
-        uses: PyO3/maturin-action@v1
+        uses: PyO3/maturin-action@5f8a1b3b0aad13193f46c9131f9b9e663def8ce5  #v1.46.0
         with:
           target: ${{ matrix.platform.target }}
           manylinux: musllinux_1_2
           args: --release --out dist --features python-bindings
 
-      - uses: uraimo/run-on-arch-action@v2.5.0
+      - uses: uraimo/run-on-arch-action@5397f9e30a9b62422f302092631c99ae1effcd9e  #v2.8.1
         name: Install built wheel
         with:
           arch: ${{ matrix.platform.arch }}
@@ -323,9 +334,9 @@ jobs:
             python3 -c "import ezkl"
 
       - name: Upload wheels
-        uses: actions/upload-artifact@v3
+        uses: actions/upload-artifact@65c4c4a1ddee5b72f698fdd19549f0f0fb45cf08  #v4.6.0
         with:
-          name: wheels
+          name: dist-musllinux-${{ matrix.platform.target }}
           path: dist
 
   pypi-publish:
@@ -334,44 +345,43 @@ jobs:
     permissions:
       id-token: write
     if: "startsWith(github.ref, 'refs/tags/')"
-    # TODO: Uncomment if linux-cross is working
-    # needs: [ macos, windows, linux, linux-cross, musllinux, musllinux-cross ]
     needs: [macos, windows, linux, musllinux, musllinux-cross]
     steps:
-      - uses: actions/download-artifact@v3
+      - uses: actions/download-artifact@fa0a91b85d4f404e444e00e005971372dc801d16  #v4.1.8
         with:
-          name: wheels
+          pattern: dist-*
+          merge-multiple: true
+          path: wheels
       - name: List Files
         run: ls -R
 
-      # Both publish steps will fail if there is no trusted publisher setup
-      # On failure the publish step will then simply continue to the next one
+      # # publishes to TestPyPI
+      # - name: Publish package distribution to TestPyPI
+      #   uses: pypa/gh-action-pypi-publish@76f52bc884231f62b9a034ebfe128415bbaabdfc  #v1.12.4
+      #   with:
+      #     repository-url: https://test.pypi.org/legacy/
+      #     packages-dir: ./
 
       # publishes to PyPI
       - name: Publish package distributions to PyPI
-        continue-on-error: true
-        uses: pypa/gh-action-pypi-publish@release/v1
+        uses: pypa/gh-action-pypi-publish@76f52bc884231f62b9a034ebfe128415bbaabdfc  #v1.12.4
         with:
-          packages-dir: ./
+          packages-dir: ./wheels
 
-      # publishes to TestPyPI
-      - name: Publish package distribution to TestPyPI
-        continue-on-error: true
-        uses: pypa/gh-action-pypi-publish@release/v1
-        with:
-          repository-url: https://test.pypi.org/legacy/
-          packages-dir: ./
 
   doc-publish:
+    permissions:
+      contents: read
     name: Trigger ReadTheDocs Build
     runs-on: ubuntu-latest
     needs: pypi-publish
     steps:
       - uses: actions/checkout@v4
-
+        with:
+          persist-credentials: false
       - name: Trigger RTDs build
         uses: dfm/rtds-action@v1
         with:
           webhook_url: ${{ secrets.RTDS_WEBHOOK_URL }}
           webhook_token: ${{ secrets.RTDS_WEBHOOK_TOKEN }}
-          commit_ref: ${{ github.ref_name }}
+          commit_ref: ${{ github.ref_name }}

.github/workflows/release.yml

@@ -10,6 +10,9 @@ on:
       - "*"
 jobs:
   create-release:
+    permissions:
+      contents: read
+      packages: write
     name: create-release
     runs-on: ubuntu-22.04
     if: startsWith(github.ref, 'refs/tags/')
@@ -27,12 +30,15 @@ jobs:
 
       - name: Create Github Release
         id: create-release
-        uses: softprops/action-gh-release@v1
+        uses: softprops/action-gh-release@c95fe1489396fe8a9eb87c0abf8aa5b2ef267fda  #v2.2.1
         with:
           token: ${{ secrets.RELEASE_TOKEN }}
           tag_name: ${{ env.EZKL_VERSION }}
 
   build-release-gpu:
+    permissions:
+      contents: read
+      packages: write
     name: build-release-gpu
     needs: ["create-release"]
     runs-on: GPU
@@ -43,13 +49,16 @@ jobs:
       RUST_BACKTRACE: 1
       PCRE2_SYS_STATIC: 1
     steps:
-      - uses: actions-rs/toolchain@v1
+      - uses: actions-rs/toolchain@b2417cde72dcf67f306c0ae8e0828a81bf0b189f  #v1.0.6
         with:
-          toolchain: nightly-2024-02-06
+          toolchain: nightly-2024-07-18
           override: true
           components: rustfmt, clippy
       - name: Checkout repo
-        uses: actions/checkout@v4
+        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683  #v4.2.2
+        with:
+          persist-credentials: false
+
 
       - name: Get release version from tag
         shell: bash
@@ -81,7 +90,7 @@ jobs:
           echo "ASSET=build-artifacts/ezkl-linux-gpu.tar.gz" >> $GITHUB_ENV
 
       - name: Upload release archive
-        uses: actions/upload-release-asset@v1.0.2
+        uses: actions/upload-release-asset@e8f9f06c4b078e705bd2ea027f0926603fc9b4d5  #v1.0.2
         env:
           GITHUB_TOKEN: ${{ secrets.RELEASE_TOKEN }}
         with:
@@ -91,6 +100,10 @@ jobs:
           asset_content_type: application/octet-stream
 
   build-release:
+    permissions:
+      contents: read
+      packages: write
+      issues: write
     name: build-release
     needs: ["create-release"]
     runs-on: ${{ matrix.os }}
@@ -106,32 +119,34 @@ jobs:
         include:
           - build: windows-msvc
             os: windows-latest
-            rust: nightly-2024-02-06
+            rust: nightly-2024-07-18
             target: x86_64-pc-windows-msvc
           - build: macos
             os: macos-13
-            rust: nightly-2024-02-06
+            rust: nightly-2024-07-18
             target: x86_64-apple-darwin
           - build: macos-aarch64
             os: macos-13
-            rust: nightly-2024-02-06
+            rust: nightly-2024-07-18
             target: aarch64-apple-darwin
           - build: linux-musl
             os: ubuntu-22.04
-            rust: nightly-2024-02-06
+            rust: nightly-2024-07-18
             target: x86_64-unknown-linux-musl
           - build: linux-gnu
             os: ubuntu-22.04
-            rust: nightly-2024-02-06
+            rust: nightly-2024-07-18
             target: x86_64-unknown-linux-gnu
           - build: linux-aarch64
             os: ubuntu-22.04
-            rust: nightly-2024-02-06
+            rust: nightly-2024-07-18
             target: aarch64-unknown-linux-gnu
 
     steps:
       - name: Checkout repo
-        uses: actions/checkout@v4
+        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683  #v4.2.2
+        with:
+          persist-credentials: false
 
       - name: Get release version from tag
         shell: bash
@@ -155,7 +170,7 @@ jobs:
           fi
 
       - name: Install Rust
-        uses: dtolnay/rust-toolchain@nightly
+        uses: dtolnay/rust-toolchain@4f94fbe7e03939b0e674bcc9ca609a16088f63ff  #nightly branch, TODO: update when required
         with:
           target: ${{ matrix.target }}
 
@@ -181,9 +196,18 @@ jobs:
           echo "target flag is: ${{ env.TARGET_FLAGS }}"
           echo "target dir is: ${{ env.TARGET_DIR }}"
 
-      - name: Build release binary
+      - name: Build release binary (no asm or metal)
+        if: matrix.build != 'linux-gnu' && matrix.build != 'macos-aarch64'
         run: ${{ env.CARGO }} build --release ${{ env.TARGET_FLAGS }} -Z sparse-registry
 
+      - name: Build release binary (asm)
+        if: matrix.build == 'linux-gnu'
+        run: ${{ env.CARGO }} build --release ${{ env.TARGET_FLAGS }} -Z sparse-registry --features asm
+
+      - name: Build release binary (metal)
+        if: matrix.build == 'macos-aarch64'
+        run: ${{ env.CARGO }} build --release ${{ env.TARGET_FLAGS }} -Z sparse-registry --features macos-metal
+
       - name: Strip release binary
         if: matrix.build != 'windows-msvc' && matrix.build != 'linux-aarch64'
         run: strip "target/${{ matrix.target }}/release/ezkl"
@@ -209,7 +233,7 @@ jobs:
           echo "ASSET=build-artifacts/ezkl-win.zip" >> $GITHUB_ENV
 
       - name: Upload release archive
-        uses: actions/upload-release-asset@v1.0.2
+        uses: actions/upload-release-asset@e8f9f06c4b078e705bd2ea027f0926603fc9b4d5  #v1.0.2
         env:
           GITHUB_TOKEN: ${{ secrets.RELEASE_TOKEN }}
         with:

.github/workflows/static-analysis.yml

@@ -0,0 +1,32 @@
+name: Static Analysis
+
+on:
+  push:
+    branches: [ main ]
+  pull_request:
+    branches: [ main ]
+
+jobs:
+  analyze:
+    permissions:
+      contents: read
+    runs-on: ubuntu-latest
+    steps:
+      - uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683  #v4.2.2
+        with:
+          persist-credentials: false
+      - uses: actions-rs/toolchain@b2417cde72dcf67f306c0ae8e0828a81bf0b189f  #v1.0.6
+        with:
+          toolchain: nightly-2024-07-18
+          override: true
+          components: rustfmt, clippy
+
+      # Run Zizmor static analysis
+
+      - name: Install Zizmor
+        run: cargo install --locked zizmor
+
+      - name: Run Zizmor Analysis
+        run: zizmor .
+
+

.github/workflows/swift-pm.yml

@@ -0,0 +1,134 @@
+name: Build and Publish EZKL iOS SPM package
+
+on:
+  push:
+    tags:
+      # Only support SemVer versioning tags
+      - 'v[0-9]+.[0-9]+.[0-9]+'
+      - '[0-9]+.[0-9]+.[0-9]+'
+
+jobs:
+  build-and-update:
+    permissions:
+      contents: read
+      packages: write
+    runs-on: macos-latest
+    env:
+      EZKL_SWIFT_PACKAGE_REPO: github.com/zkonduit/ezkl-swift-package.git
+      RELEASE_TAG: ${{ github.ref_name }}
+
+    steps:
+      - name: Checkout EZKL
+        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683  #v4.2.2
+        with:
+          persist-credentials: false
+
+      - name: Extract TAG from github.ref_name
+        run: |
+          # github.ref_name is provided by GitHub Actions and contains the tag name directly.
+          TAG="${RELEASE_TAG}"
+          echo "Original TAG: $TAG"
+          # Remove leading 'v' if present to match the Swift Package Manager version format.
+          NEW_TAG=${TAG#v}
+          echo "Stripped TAG: $NEW_TAG"
+          echo "TAG=$NEW_TAG" >> $GITHUB_ENV
+
+      - name: Install Rust (nightly)
+        uses: actions-rs/toolchain@b2417cde72dcf67f306c0ae8e0828a81bf0b189f  #v1.0.6
+        with:
+          toolchain: nightly
+          override: true
+
+      - name: Build EzklCoreBindings
+        run: CONFIGURATION=release cargo run --bin ios_gen_bindings --features "ios-bindings uuid camino uniffi_bindgen" --no-default-features
+
+      - name: Clone ezkl-swift-package repository
+        run: |
+          git clone https://${{ env.EZKL_SWIFT_PACKAGE_REPO }}
+
+      - name: Copy EzklCoreBindings
+        run: |
+          rm -rf ezkl-swift-package/Sources/EzklCoreBindings
+          cp -r build/EzklCoreBindings ezkl-swift-package/Sources/
+
+      - name: Copy Test Files
+        run: |
+          rm -rf ezkl-swift-package/Tests/EzklAssets/
+          mkdir -p ezkl-swift-package/Tests/EzklAssets/
+          cp tests/assets/kzg ezkl-swift-package/Tests/EzklAssets/kzg.srs
+          cp tests/assets/input.json ezkl-swift-package/Tests/EzklAssets/input.json
+          cp tests/assets/model.compiled ezkl-swift-package/Tests/EzklAssets/network.ezkl
+          cp tests/assets/settings.json ezkl-swift-package/Tests/EzklAssets/settings.json
+
+      - name: Check for changes
+        id: check_changes
+        run: |
+          cd ezkl-swift-package
+          if git diff --quiet Sources/EzklCoreBindings Tests/EzklAssets; then
+            echo "no_changes=true" >> $GITHUB_OUTPUT
+          else
+            echo "no_changes=false" >> $GITHUB_OUTPUT
+          fi
+
+      - name: Set up Xcode environment
+        if: steps.check_changes.outputs.no_changes == 'false'
+        run: |
+          sudo xcode-select -s /Applications/Xcode.app/Contents/Developer
+          sudo xcodebuild -license accept
+
+      - name: Run Package Tests
+        if: steps.check_changes.outputs.no_changes == 'false'
+        run: |
+          cd ezkl-swift-package
+          xcodebuild test \
+            -scheme EzklPackage \
+            -destination 'platform=iOS Simulator,name=iPhone 15 Pro,OS=17.5' \
+            -resultBundlePath ../testResults
+
+      - name: Run Example App Tests
+        if: steps.check_changes.outputs.no_changes == 'false'
+        run: |
+          cd ezkl-swift-package/Example
+          xcodebuild test \
+            -project Example.xcodeproj \
+            -scheme EzklApp \
+            -destination 'platform=iOS Simulator,name=iPhone 15 Pro,OS=17.5' \
+            -parallel-testing-enabled NO \
+            -resultBundlePath ../../exampleTestResults \
+            -skip-testing:EzklAppUITests/EzklAppUITests/testButtonClicksInOrder
+
+      - name: Setup Git
+        run: |
+          cd ezkl-swift-package
+          git config user.name "GitHub Action"
+          git config user.email "action@github.com"
+          git remote set-url origin https://zkonduit:${EZKL_SWIFT_PACKAGE_REPO_TOKEN}@${{ env.EZKL_SWIFT_PACKAGE_REPO }}
+        env:
+          EZKL_SWIFT_PACKAGE_REPO_TOKEN: ${{ secrets.EZKL_PORTER_TOKEN }}
+
+      - name: Commit and Push Changes
+        if: steps.check_changes.outputs.no_changes == 'false'
+        run: |
+          cd ezkl-swift-package
+          git add Sources/EzklCoreBindings Tests/EzklAssets
+          git commit -m "Automatically updated EzklCoreBindings for EZKL"
+          if ! git push origin; then
+            echo "::error::Failed to push changes to ${{ env.EZKL_SWIFT_PACKAGE_REPO }}. Please ensure that EZKL_PORTER_TOKEN has the correct permissions."
+            exit 1
+          fi
+
+      - name: Tag the latest commit
+        run: |
+          cd ezkl-swift-package
+          source $GITHUB_ENV
+          # Tag the latest commit on the current branch
+          if git rev-parse "$TAG" >/dev/null 2>&1; then
+            echo "Tag $TAG already exists locally. Skipping tag creation."
+          else
+            git tag "$TAG"
+          fi
+
+          if ! git push origin "$TAG"; then
+            echo "::error::Failed to push tag '$TAG' to ${{ env.EZKL_SWIFT_PACKAGE_REPO }}. Please ensure EZKL_PORTER_TOKEN has correct permissions."
+            exit 1
+          fi

.github/workflows/tagging.yml

@@ -11,10 +11,12 @@ jobs:
       contents: write
 
     steps:
-      - uses: actions/checkout@v4
+      - uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683  #v4.2.2
+        with:
+          persist-credentials: false
       - name: Bump version and push tag
         id: tag_version
-        uses: mathieudutour/github-tag-action@v6.2
+        uses: mathieudutour/github-tag-action@a22cf08638b34d5badda920f9daf6e72c477b07b  #v6.2
         with:
           github_token: ${{ secrets.GITHUB_TOKEN }}
 
@@ -44,7 +46,7 @@ jobs:
           git tag $RELEASE_TAG
 
       - name: Push changes
-        uses: ad-m/github-push-action@master
+        uses: ad-m/github-push-action@77c5b412c50b723d2a4fbc6d71fb5723bcd439aa  #master
         env:
           RELEASE_TAG: ${{ steps.tag_version.outputs.new_tag }}
         with:

.gitignore

@@ -27,7 +27,6 @@ __pycache__/
 *.pyc
 *.pyo
 *.py[cod]
-bin/
 build/
 develop-eggs/
 dist/
@@ -46,7 +45,7 @@ var/
 node_modules
 /dist
 timingData.json
-!tests/wasm/pk.key
-!tests/wasm/vk.key
+!tests/assets/pk.key
+!tests/assets/vk.key
 docs/python/build
-!tests/wasm/vk_aggr.key
+!tests/assets/vk_aggr.key

Cargo.toml

@@ -4,6 +4,7 @@ cargo-features = ["profile-rustflags"]
 name = "ezkl"
 version = "0.0.0"
 edition = "2021"
+default-run = "ezkl"
 
 # See more keys and their definitions at https://doc.rust-lang.org/cargo/reference/manifest.html
 
@@ -11,86 +12,106 @@ edition = "2021"
 # Name to be imported within python
 # Example: import ezkl
 name = "ezkl"
-crate-type = ["cdylib", "rlib"]
+crate-type = ["cdylib", "rlib", "staticlib"]
 
 
 [dependencies]
-halo2_gadgets = { git = "https://github.com/zkonduit/halo2", branch = "ac/optional-selector-poly" }
-halo2_proofs = { git = "https://github.com/zkonduit/halo2", branch = "ac/optional-selector-poly" }
-halo2curves = { git = "https://github.com/privacy-scaling-explorations/halo2curves", rev = "9fff22c", features = [
+halo2_gadgets = { git = "https://github.com/zkonduit/halo2" }
+halo2curves = { git = "https://github.com/privacy-scaling-explorations/halo2curves", rev = "b753a832e92d5c86c5c997327a9cf9de86a18851", features = [
     "derive_serde",
 ] }
-rand = { version = "0.8", default_features = false }
-itertools = { version = "0.10.3", default_features = false }
-clap = { version = "4.5.3", features = ["derive"] }
-clap_complete = "4.5.2"
-serde = { version = "1.0.126", features = ["derive"], optional = true }
-serde_json = { version = "1.0.97", default_features = false, features = [
-    "float_roundtrip",
-    "raw_value",
-], optional = true }
-log = { version = "0.4.17", default_features = false, optional = true }
-thiserror = { version = "1.0.38", default_features = false }
-hex = { version = "0.4.3", default_features = false }
+halo2_proofs = { git = "https://github.com/zkonduit/halo2", package = "halo2_proofs", features = [
+    "circuit-params",
+] }
+rand = { version = "0.8", default-features = false }
+itertools = { version = "0.10.3", default-features = false }
+clap = { version = "4.5.3", features = ["derive"], optional = true }
+serde = { version = "1.0.126", features = ["derive"] }
+clap_complete = { version = "4.5.2", optional = true }
+log = { version = "0.4.17", default-features = false }
+thiserror = { version = "1.0.38", default-features = false }
+hex = { version = "0.4.3", default-features = false }
 halo2_wrong_ecc = { git = "https://github.com/zkonduit/halo2wrong", branch = "ac/chunked-mv-lookup", package = "ecc" }
 snark-verifier = { git = "https://github.com/zkonduit/snark-verifier", branch = "ac/chunked-mv-lookup", features = [
     "derive_serde",
 ] }
-halo2_solidity_verifier = { git = "https://github.com/alexander-camuto/halo2-solidity-verifier", branch = "main" }
-maybe-rayon = { version = "0.1.1", default_features = false }
-bincode = { version = "1.3.3", default_features = false }
+halo2_solidity_verifier = { git = "https://github.com/alexander-camuto/halo2-solidity-verifier", optional = true }
+maybe-rayon = { version = "0.1.1", default-features = false }
+bincode = { version = "1.3.3", default-features = false }
 unzip-n = "0.1.2"
 num = "0.4.1"
-portable-atomic = "1.6.0"
-tosubcommand = { git = "https://github.com/zkonduit/enum_to_subcommand", package = "tosubcommand" }
-metal = { git = "https://github.com/gfx-rs/metal-rs", optional = true }
-semver = "1.0.22"
+tosubcommand = { git = "https://github.com/zkonduit/enum_to_subcommand", package = "tosubcommand", optional = true }
+semver = { version = "1.0.22", optional = true }
+
+[target.'cfg(not(target_arch = "wasm32"))'.dependencies]
+serde_json = { version = "1.0.97", features = ["float_roundtrip", "raw_value"] }
 
 # evm related deps
-[target.'cfg(not(target_arch = "wasm32"))'.dependencies]
-alloy = { git = "https://github.com/alloy-rs/alloy", version = "0.1.0", rev="5fbf57bac99edef9d8475190109a7ea9fb7e5e83", features = ["provider-http", "signers", "contract", "rpc-types-eth", "signer-wallet", "node-bindings"] }
-foundry-compilers = {version = "0.4.1", features = ["svm-solc"]}
-ethabi = "18"
-indicatif = { version = "0.17.5", features = ["rayon"] }
-gag = { version = "1.0.0", default_features = false }
+alloy = { git = "https://github.com/alloy-rs/alloy", version = "0.1.0", rev = "5fbf57bac99edef9d8475190109a7ea9fb7e5e83", features = [
+    "provider-http",
+    "signers",
+    "contract",
+    "rpc-types-eth",
+    "signer-wallet",
+    "node-bindings",
+
+], optional = true }
+foundry-compilers = { version = "0.4.1", features = [
+    "svm-solc",
+], optional = true }
+ethabi = { version = "18", optional = true }
+indicatif = { version = "0.17.5", features = ["rayon"], optional = true }
+gag = { version = "1.0.0", default-features = false, optional = true }
 instant = { version = "0.1" }
 reqwest = { version = "0.12.4", default-features = false, features = [
     "default-tls",
     "multipart",
     "stream",
-] }
-openssl = { version = "0.10.55", features = ["vendored"] }
-tokio-postgres = "0.7.10"
-pg_bigdecimal = "0.1.5"
-lazy_static = "1.4.0"
-colored_json = { version = "3.0.1", default_features = false, optional = true }
-regex = { version = "1", default_features = false }
-tokio = { version = "1.35", default_features = false, features = [
+], optional = true }
+openssl = { version = "0.10.55", features = ["vendored"], optional = true }
+tokio-postgres = { version = "0.7.10", optional = true }
+pg_bigdecimal = { version = "0.1.5", optional = true }
+lazy_static = { version = "1.4.0", optional = true }
+colored_json = { version = "3.0.1", default-features = false, optional = true }
+tokio = { version = "1.35.0", default-features = false, features = [
     "macros",
-    "rt-multi-thread"
-] }
-pyo3 = { version = "0.21.2", features = [
+    "rt-multi-thread",
+], optional = true }
+pyo3 = { version = "0.23.2", features = [
     "extension-module",
     "abi3-py37",
     "macros",
-], default_features = false, optional = true }
-pyo3-asyncio = { git = "https://github.com/jopemachine/pyo3-asyncio/", branch="migration-pyo3-0.21", features = [
+], default-features = false, optional = true }
+pyo3-async-runtimes = { git = "https://github.com/PyO3/pyo3-async-runtimes", version = "0.23.0", features = [
     "attributes",
     "tokio-runtime",
-], default_features = false, optional = true }
-pyo3-log = { version = "0.10.0", default_features = false, optional = true }
-tract-onnx = { git = "https://github.com/sonos/tract/", rev = "7bf303b2ae9bddd5fa6951ae95848c0d52fb7f50", default_features = false, optional = true }
+], default-features = false, optional = true }
+pyo3-log = { version = "0.12.0", default-features = false, optional = true }
+tract-onnx = { git = "https://github.com/sonos/tract/", rev = "37132e0397d0a73e5bd3a8615d932dabe44f6736", default-features = false, optional = true }
 tabled = { version = "0.12.0", optional = true }
 objc = { version = "0.2.4", optional = true }
+mimalloc = { version = "0.1", optional = true }
+pyo3-stub-gen = { version = "0.6.0", optional = true }
 
+# universal bindings
+uniffi = { version = "=0.28.0", optional = true }
+getrandom = { version = "0.2.8", optional = true }
+uniffi_bindgen = { version = "=0.28.0", optional = true }
+camino = { version = "^1.1", optional = true }
+uuid = { version = "1.10.0", features = ["v4"], optional = true }
 
 [target.'cfg(not(all(target_arch = "wasm32", target_os = "unknown")))'.dependencies]
-colored = { version = "2.0.0", default_features = false, optional = true }
-env_logger = { version = "0.10.0", default_features = false, optional = true }
-chrono = "0.4.31"
-sha256 = "1.4.0"
+colored = { version = "2.0.0", default-features = false, optional = true }
+env_logger = { version = "0.10.0", default-features = false, optional = true }
+chrono = { version = "0.4.31", optional = true }
+sha256 = { version = "1.4.0", optional = true }
+
 
 [target.'cfg(target_arch = "wasm32")'.dependencies]
+serde_json = { version = "1.0.97", default-features = false, features = [
+    "float_roundtrip",
+    "raw_value",
+] }
 getrandom = { version = "0.2.8", features = ["js"] }
 instant = { version = "0.1", features = ["wasm-bindgen", "inaccurate"] }
 
@@ -106,6 +127,10 @@ wasm-bindgen-console-logger = "0.1.1"
 [target.'cfg(not(all(target_arch = "wasm32", target_os = "unknown")))'.dev-dependencies]
 criterion = { version = "0.5.1", features = ["html_reports"] }
 
+
+[build-dependencies]
+uniffi = { version = "0.28", features = ["build"], optional = true }
+
 [dev-dependencies]
 tempfile = "3.3.0"
 lazy_static = "1.4.0"
@@ -119,6 +144,10 @@ shellexpand = "3.1.0"
 runner = 'wasm-bindgen-test-runner'
 
 
+[[bench]]
+name = "zero_finder"
+harness = false
+
 [[bench]]
 name = "accum_dot"
 harness = false
@@ -157,16 +186,20 @@ harness = false
 
 
 [[bench]]
-name = "relu"
+name = "sigmoid"
 harness = false
 
 [[bench]]
-name = "accum_matmul_relu"
+name = "relu_lookupless"
+harness = false
+
+[[bench]]
+name = "accum_matmul_sigmoid"
 harness = false
 
 
 [[bench]]
-name = "accum_matmul_relu_overflow"
+name = "accum_matmul_sigmoid_overflow"
 harness = false
 
 [[bin]]
@@ -175,42 +208,97 @@ test = false
 bench = false
 required-features = ["ezkl"]
 
+[[bin]]
+name = "ios_gen_bindings"
+required-features = ["ios-bindings", "uuid", "camino", "uniffi_bindgen"]
+
+[[bin]]
+name = "py_stub_gen"
+required-features = ["python-bindings"]
+
 [features]
 web = ["wasm-bindgen-rayon"]
-default = ["ezkl", "mv-lookup", "no-banner", "parallel-poly-read"]
+default = [
+    "ezkl",
+    "mv-lookup",
+    "precompute-coset",
+    "no-banner",
+    "parallel-poly-read",
+]
 onnx = ["dep:tract-onnx"]
-python-bindings = ["pyo3", "pyo3-log", "pyo3-asyncio"]
+python-bindings = ["pyo3", "pyo3-log", "pyo3-async-runtimes", "pyo3-stub-gen"]
+ios-bindings = ["mv-lookup", "precompute-coset", "parallel-poly-read", "uniffi"]
+ios-bindings-test = ["ios-bindings", "uniffi/bindgen-tests"]
 ezkl = [
     "onnx",
-    "serde",
-    "serde_json",
-    "log",
-    "colored",
-    "env_logger",
+    "dep:colored",
+    "dep:env_logger",
     "tabled/color",
+    "serde_json/std",
     "colored_json",
-    "halo2_proofs/circuit-params",
+    "dep:alloy",
+    "dep:foundry-compilers",
+    "dep:ethabi",
+    "dep:indicatif",
+    "dep:gag",
+    "dep:reqwest",
+    "dep:tokio-postgres",
+    "dep:pg_bigdecimal",
+    "dep:lazy_static",
+    "dep:tokio",
+    "dep:openssl",
+    "dep:mimalloc",
+    "dep:chrono",
+    "dep:sha256",
+    "dep:clap_complete",
+    "dep:halo2_solidity_verifier",
+    "dep:semver",
+    "dep:clap",
+    "dep:tosubcommand",
+]
+parallel-poly-read = [
+    "halo2_proofs/circuit-params",
+    "halo2_proofs/parallel-poly-read",
 ]
-parallel-poly-read = ["halo2_proofs/parallel-poly-read"]
 mv-lookup = [
     "halo2_proofs/mv-lookup",
     "snark-verifier/mv-lookup",
     "halo2_solidity_verifier/mv-lookup",
 ]
+asm = ["halo2curves/asm", "halo2_proofs/asm"]
+precompute-coset = ["halo2_proofs/precompute-coset"]
 det-prove = []
 icicle = ["halo2_proofs/icicle_gpu"]
 empty-cmd = []
 no-banner = []
 no-update = []
-metal = ["dep:metal", "dep:objc"]
-
-# icicle patch to 0.1.0 if feature icicle is enabled
-[patch.'https://github.com/ingonyama-zk/icicle']
-icicle = { git = "https://github.com/ingonyama-zk/icicle?rev=45b00fb", package = "icicle", branch = "fix/vhnat/ezkl-build-fix" }
+macos-metal = ["halo2_proofs/macos"]
+ios-metal = ["halo2_proofs/ios"]
 
 [patch.'https://github.com/zkonduit/halo2']
-halo2_proofs = { git = "https://github.com/zkonduit/halo2#8cfca221f53069a0374687654882b99e729041d7", package = "halo2_proofs" }
+halo2_proofs = { git = "https://github.com/zkonduit/halo2#f441c920be45f8f05d2c06a173d82e8885a5ed4d", package = "halo2_proofs" }
 
+[patch.'https://github.com/zkonduit/halo2#0654e92bdf725fd44d849bfef3643870a8c7d50b']
+halo2_proofs = { git = "https://github.com/zkonduit/halo2#f441c920be45f8f05d2c06a173d82e8885a5ed4d", package = "halo2_proofs" }
+
+
+[patch.crates-io]
+uniffi_testing = { git = "https://github.com/ElusAegis/uniffi-rs", branch = "feat/testing-feature-build-fix" }
 
 [profile.release]
 rustflags = ["-C", "relocation-model=pic"]
+lto = "fat"
+codegen-units = 1
+#panic = "abort"
+
+
+[profile.test-runs]
+inherits = "dev"
+opt-level = 3
+
+[package.metadata.wasm-pack.profile.release]
+wasm-opt = [
+    "-O4",
+    "--flexible-inline-max-function-size",
+    "4294967295",
+]

README.md

@@ -150,6 +150,13 @@ Ezkl is unaudited, beta software undergoing rapid development. There may be bugs
 
 > NOTE: Because operations are quantized when they are converted from an onnx file to a zk-circuit, outputs in python and ezkl may differ slightly. 
 
+
+### Advanced security topics 
+
+Check out `docs/advanced_security` for more advanced information on potential threat vectors. 
+
+
+
 ### no warranty
 
 Copyright (c) 2024 Zkonduit Inc. THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,

abis/DataAttestationSingle.json

@@ -0,0 +1,147 @@
+[
+    {
+        "inputs": [
+            {
+                "internalType": "address",
+                "name": "_contractAddresses",
+                "type": "address"
+            },
+            {
+                "internalType": "bytes",
+                "name": "_callData",
+                "type": "bytes"
+            },
+            {
+                "internalType": "uint256",
+                "name": "_decimals",
+                "type": "uint256"
+            },
+            {
+                "internalType": "uint256[]",
+                "name": "_scales",
+                "type": "uint256[]"
+            },
+            {
+                "internalType": "uint8",
+                "name": "_instanceOffset",
+                "type": "uint8"
+            },
+            {
+                "internalType": "address",
+                "name": "_admin",
+                "type": "address"
+            }
+        ],
+        "stateMutability": "nonpayable",
+        "type": "constructor"
+    },
+    {
+        "inputs": [],
+        "name": "accountCall",
+        "outputs": [
+            {
+                "internalType": "address",
+                "name": "contractAddress",
+                "type": "address"
+            },
+            {
+                "internalType": "bytes",
+                "name": "callData",
+                "type": "bytes"
+            },
+            {
+                "internalType": "uint256",
+                "name": "decimals",
+                "type": "uint256"
+            }
+        ],
+        "stateMutability": "view",
+        "type": "function"
+    },
+    {
+        "inputs": [],
+        "name": "admin",
+        "outputs": [
+            {
+                "internalType": "address",
+                "name": "",
+                "type": "address"
+            }
+        ],
+        "stateMutability": "view",
+        "type": "function"
+    },
+    {
+        "inputs": [],
+        "name": "instanceOffset",
+        "outputs": [
+            {
+                "internalType": "uint8",
+                "name": "",
+                "type": "uint8"
+            }
+        ],
+        "stateMutability": "view",
+        "type": "function"
+    },
+    {
+        "inputs": [
+            {
+                "internalType": "address",
+                "name": "_contractAddresses",
+                "type": "address"
+            },
+            {
+                "internalType": "bytes",
+                "name": "_callData",
+                "type": "bytes"
+            },
+            {
+                "internalType": "uint256",
+                "name": "_decimals",
+                "type": "uint256"
+            }
+        ],
+        "name": "updateAccountCalls",
+        "outputs": [],
+        "stateMutability": "nonpayable",
+        "type": "function"
+    },
+    {
+        "inputs": [
+            {
+                "internalType": "address",
+                "name": "_admin",
+                "type": "address"
+            }
+        ],
+        "name": "updateAdmin",
+        "outputs": [],
+        "stateMutability": "nonpayable",
+        "type": "function"
+    },
+    {
+        "inputs": [
+            {
+                "internalType": "address",
+                "name": "verifier",
+                "type": "address"
+            },
+            {
+                "internalType": "bytes",
+                "name": "encoded",
+                "type": "bytes"
+            }
+        ],
+        "name": "verifyWithDataAttestation",
+        "outputs": [
+            {
+                "internalType": "bool",
+                "name": "",
+                "type": "bool"
+            }
+        ],
+        "stateMutability": "view",
+        "type": "function"
+    }
+]

abis/QuantizeData.json

@@ -1,4 +1,23 @@
 [
+	{
+		"inputs": [
+			{
+				"internalType": "int256[]",
+				"name": "quantized_data",
+				"type": "int256[]"
+			}
+		],
+		"name": "check_is_valid_field_element",
+		"outputs": [
+			{
+				"internalType": "uint256[]",
+				"name": "output",
+				"type": "uint256[]"
+			}
+		],
+		"stateMutability": "pure",
+		"type": "function"
+	},
 	{
 		"inputs": [
 			{
@@ -17,12 +36,41 @@
 				"type": "uint256[]"
 			}
 		],
-		"name": "quantize_data",
+		"name": "quantize_data_multi",
 		"outputs": [
 			{
-				"internalType": "int64[]",
+				"internalType": "int256[]",
 				"name": "quantized_data",
-				"type": "int64[]"
+				"type": "int256[]"
+			}
+		],
+		"stateMutability": "pure",
+		"type": "function"
+	},
+	{
+		"inputs": [
+			{
+				"internalType": "bytes",
+				"name": "data",
+				"type": "bytes"
+			},
+			{
+				"internalType": "uint256",
+				"name": "decimals",
+				"type": "uint256"
+			},
+			{
+				"internalType": "uint256[]",
+				"name": "scales",
+				"type": "uint256[]"
+			}
+		],
+		"name": "quantize_data_single",
+		"outputs": [
+			{
+				"internalType": "int256[]",
+				"name": "quantized_data",
+				"type": "int256[]"
 			}
 		],
 		"stateMutability": "pure",

benches/accum_conv.rs

@@ -64,7 +64,7 @@ impl Circuit<Fr> for MyCircuit {
         layouter.assign_region(
             || "",
             |region| {
-                let mut region = region::RegionCtx::new(region, 0, 1);
+                let mut region = region::RegionCtx::new(region, 0, 1, 1024, 2);
                 config
                     .layout(
                         &mut region,
@@ -72,6 +72,9 @@ impl Circuit<Fr> for MyCircuit {
                         Box::new(PolyOp::Conv {
                             padding: vec![(0, 0)],
                             stride: vec![1; 2],
+                            group: 1,
+                            data_format: DataFormat::NCHW,
+                            kernel_format: KernelFormat::OIHW,
                         }),
                     )
                     .unwrap();

benches/accum_dot.rs

@@ -55,7 +55,7 @@ impl Circuit<Fr> for MyCircuit {
         layouter.assign_region(
             || "",
             |region| {
-                let mut region = region::RegionCtx::new(region, 0, 1);
+                let mut region = region::RegionCtx::new(region, 0, 1, 1024, 2);
                 config
                     .layout(
                         &mut region,

benches/accum_einsum_matmul.rs

@@ -57,7 +57,7 @@ impl Circuit<Fr> for MyCircuit {
         layouter.assign_region(
             || "",
             |region| {
-                let mut region = region::RegionCtx::new(region, 0, 1);
+                let mut region = region::RegionCtx::new(region, 0, 1, 1024, 2);
                 config
                     .layout(
                         &mut region,

benches/accum_matmul_sigmoid.rs

@@ -57,7 +57,15 @@ impl Circuit<Fr> for MyCircuit {
 
         // sets up a new relu table
         base_config
-            .configure_lookup(cs, &b, &output, &a, BITS, K, &LookupOp::ReLU)
+            .configure_lookup(
+                cs,
+                &b,
+                &output,
+                &a,
+                BITS,
+                K,
+                &LookupOp::Sigmoid { scale: 1.0.into() },
+            )
             .unwrap();
 
         MyConfig { base_config }
@@ -75,14 +83,18 @@ impl Circuit<Fr> for MyCircuit {
                 let op = PolyOp::Einsum {
                     equation: "ij,jk->ik".to_string(),
                 };
-                let mut region = region::RegionCtx::new(region, 0, 1);
+                let mut region = region::RegionCtx::new(region, 0, 1, 1024, 2);
                 let output = config
                     .base_config
                     .layout(&mut region, &self.inputs, Box::new(op))
                     .unwrap();
                 let _output = config
                     .base_config
-                    .layout(&mut region, &[output.unwrap()], Box::new(LookupOp::ReLU))
+                    .layout(
+                        &mut region,
+                        &[output.unwrap()],
+                        Box::new(LookupOp::Sigmoid { scale: 1.0.into() }),
+                    )
                     .unwrap();
                 Ok(())
             },

benches/accum_matmul_sigmoid_overflow.rs

@@ -58,7 +58,15 @@ impl Circuit<Fr> for MyCircuit {
 
         // sets up a new relu table
         base_config
-            .configure_lookup(cs, &b, &output, &a, BITS, k, &LookupOp::ReLU)
+            .configure_lookup(
+                cs,
+                &b,
+                &output,
+                &a,
+                BITS,
+                k,
+                &LookupOp::Sigmoid { scale: 1.0.into() },
+            )
             .unwrap();
 
         MyConfig { base_config }
@@ -76,14 +84,18 @@ impl Circuit<Fr> for MyCircuit {
                 let op = PolyOp::Einsum {
                     equation: "ij,jk->ik".to_string(),
                 };
-                let mut region = region::RegionCtx::new(region, 0, 1);
+                let mut region = region::RegionCtx::new(region, 0, 1, 1024, 2);
                 let output = config
                     .base_config
                     .layout(&mut region, &self.inputs, Box::new(op))
                     .unwrap();
                 let _output = config
                     .base_config
-                    .layout(&mut region, &[output.unwrap()], Box::new(LookupOp::ReLU))
+                    .layout(
+                        &mut region,
+                        &[output.unwrap()],
+                        Box::new(LookupOp::Sigmoid { scale: 1.0.into() }),
+                    )
                     .unwrap();
                 Ok(())
             },

benches/accum_sum.rs

@@ -55,7 +55,7 @@ impl Circuit<Fr> for MyCircuit {
         layouter.assign_region(
             || "",
             |region| {
-                let mut region = region::RegionCtx::new(region, 0, 1);
+                let mut region = region::RegionCtx::new(region, 0, 1, 1024, 2);
                 config
                     .layout(
                         &mut region,

benches/accum_sumpool.rs

@@ -59,7 +59,7 @@ impl Circuit<Fr> for MyCircuit {
         layouter.assign_region(
             || "",
             |region| {
-                let mut region = region::RegionCtx::new(region, 0, 1);
+                let mut region = region::RegionCtx::new(region, 0, 1, 1024, 2);
                 config
                     .layout(
                         &mut region,
@@ -69,6 +69,7 @@ impl Circuit<Fr> for MyCircuit {
                             stride: vec![1, 1],
                             kernel_shape: vec![2, 2],
                             normalized: false,
+                            data_format: DataFormat::NCHW,
                         }),
                     )
                     .unwrap();

benches/pairwise_add.rs

@@ -55,7 +55,7 @@ impl Circuit<Fr> for MyCircuit {
         layouter.assign_region(
             || "",
             |region| {
-                let mut region = region::RegionCtx::new(region, 0, 1);
+                let mut region = region::RegionCtx::new(region, 0, 1, 1024, 2);
                 config
                     .layout(&mut region, &self.inputs, Box::new(PolyOp::Add))
                     .unwrap();

benches/pairwise_pow.rs

@@ -56,7 +56,7 @@ impl Circuit<Fr> for MyCircuit {
         layouter.assign_region(
             || "",
             |region| {
-                let mut region = RegionCtx::new(region, 0, 1);
+                let mut region = RegionCtx::new(region, 0, 1, 1024, 2);
                 config
                     .layout(&mut region, &self.inputs, Box::new(PolyOp::Pow(4)))
                     .unwrap();

benches/poseidon.rs

@@ -23,8 +23,6 @@ use halo2curves::bn256::{Bn256, Fr};
 use rand::rngs::OsRng;
 use snark_verifier::system::halo2::transcript::evm::EvmTranscript;
 
-const L: usize = 10;
-
 #[derive(Clone, Debug)]
 struct MyCircuit {
     image: ValTensor<Fr>,
@@ -40,7 +38,7 @@ impl Circuit<Fr> for MyCircuit {
     }
 
     fn configure(cs: &mut ConstraintSystem<Fr>) -> Self::Config {
-        PoseidonChip::<PoseidonSpec, POSEIDON_WIDTH, POSEIDON_RATE, 10>::configure(cs, ())
+        PoseidonChip::<PoseidonSpec, POSEIDON_WIDTH, POSEIDON_RATE>::configure(cs, ())
     }
 
     fn synthesize(
@@ -48,7 +46,7 @@ impl Circuit<Fr> for MyCircuit {
         config: Self::Config,
         mut layouter: impl Layouter<Fr>,
     ) -> Result<(), Error> {
-        let chip: PoseidonChip<PoseidonSpec, POSEIDON_WIDTH, POSEIDON_RATE, L> =
+        let chip: PoseidonChip<PoseidonSpec, POSEIDON_WIDTH, POSEIDON_RATE> =
             PoseidonChip::new(config);
         chip.layout(&mut layouter, &[self.image.clone()], 0, &mut HashMap::new())?;
         Ok(())
@@ -59,7 +57,7 @@ fn runposeidon(c: &mut Criterion) {
     let mut group = c.benchmark_group("poseidon");
 
     for size in [64, 784, 2352, 12288].iter() {
-        let k = (PoseidonChip::<PoseidonSpec, POSEIDON_WIDTH, POSEIDON_RATE, L>::num_rows(*size)
+        let k = (PoseidonChip::<PoseidonSpec, POSEIDON_WIDTH, POSEIDON_RATE>::num_rows(*size)
             as f32)
             .log2()
             .ceil() as u32;
@@ -67,7 +65,7 @@ fn runposeidon(c: &mut Criterion) {
 
         let message = (0..*size).map(|_| Fr::random(OsRng)).collect::<Vec<_>>();
         let _output =
-            PoseidonChip::<PoseidonSpec, POSEIDON_WIDTH, POSEIDON_RATE, L>::run(message.to_vec())
+            PoseidonChip::<PoseidonSpec, POSEIDON_WIDTH, POSEIDON_RATE>::run(message.to_vec())
                 .unwrap();
 
         let mut image = Tensor::from(message.into_iter().map(Value::known));

benches/relu_lookupless.rs

@@ -0,0 +1,150 @@
+use criterion::{criterion_group, criterion_main, BenchmarkId, Criterion, Throughput};
+use ezkl::circuit::poly::PolyOp;
+use ezkl::circuit::region::RegionCtx;
+use ezkl::circuit::{BaseConfig as Config, CheckMode};
+use ezkl::fieldutils::IntegerRep;
+use ezkl::pfsys::create_proof_circuit;
+use ezkl::pfsys::TranscriptType;
+use ezkl::pfsys::{create_keys, srs::gen_srs};
+use ezkl::tensor::*;
+use halo2_proofs::poly::kzg::commitment::KZGCommitmentScheme;
+use halo2_proofs::poly::kzg::multiopen::{ProverSHPLONK, VerifierSHPLONK};
+use halo2_proofs::poly::kzg::strategy::SingleStrategy;
+use halo2_proofs::{
+    circuit::{Layouter, SimpleFloorPlanner, Value},
+    plonk::{Circuit, ConstraintSystem, Error},
+};
+use halo2curves::bn256::{Bn256, Fr};
+use rand::Rng;
+use snark_verifier::system::halo2::transcript::evm::EvmTranscript;
+
+static mut LEN: usize = 4;
+const K: usize = 16;
+
+#[derive(Clone)]
+struct NLCircuit {
+    pub input: ValTensor<Fr>,
+}
+
+impl Circuit<Fr> for NLCircuit {
+    type Config = Config<Fr>;
+    type FloorPlanner = SimpleFloorPlanner;
+    type Params = ();
+
+    fn without_witnesses(&self) -> Self {
+        self.clone()
+    }
+
+    fn configure(cs: &mut ConstraintSystem<Fr>) -> Self::Config {
+        unsafe {
+            let advices = (0..3)
+                .map(|_| VarTensor::new_advice(cs, K, 1, LEN))
+                .collect::<Vec<_>>();
+
+            let mut config = Config::default();
+
+            config
+                .configure_range_check(cs, &advices[0], &advices[1], (-1, 1), K)
+                .unwrap();
+
+            config
+                .configure_range_check(cs, &advices[0], &advices[1], (0, 1023), K)
+                .unwrap();
+
+            let _constant = VarTensor::constant_cols(cs, K, LEN, false);
+
+            config
+        }
+    }
+
+    fn synthesize(
+        &self,
+        mut config: Self::Config,
+        mut layouter: impl Layouter<Fr>, // layouter is our 'write buffer' for the circuit
+    ) -> Result<(), Error> {
+        config.layout_range_checks(&mut layouter).unwrap();
+        layouter.assign_region(
+            || "",
+            |region| {
+                let mut region = RegionCtx::new(region, 0, 1, 1024, 2);
+                config
+                    .layout(
+                        &mut region,
+                        &[self.input.clone()],
+                        Box::new(PolyOp::LeakyReLU {
+                            slope: 0.0.into(),
+                            scale: 1,
+                        }),
+                    )
+                    .unwrap();
+                Ok(())
+            },
+        )?;
+        Ok(())
+    }
+}
+
+fn runrelu(c: &mut Criterion) {
+    let mut group = c.benchmark_group("relu");
+
+    let mut rng = rand::thread_rng();
+    let params = gen_srs::<KZGCommitmentScheme<_>>(17);
+    for &len in [4, 8].iter() {
+        unsafe {
+            LEN = len;
+        };
+
+        let input: Tensor<Value<Fr>> =
+            Tensor::<IntegerRep>::from((0..len).map(|_| rng.gen_range(0..10))).into();
+
+        let circuit = NLCircuit {
+            input: ValTensor::from(input.clone()),
+        };
+
+        group.throughput(Throughput::Elements(len as u64));
+        group.bench_with_input(BenchmarkId::new("pk", len), &len, |b, &_| {
+            b.iter(|| {
+                create_keys::<KZGCommitmentScheme<Bn256>, NLCircuit>(&circuit, &params, true)
+                    .unwrap();
+            });
+        });
+
+        let pk =
+            create_keys::<KZGCommitmentScheme<Bn256>, NLCircuit>(&circuit, &params, true).unwrap();
+
+        group.throughput(Throughput::Elements(len as u64));
+        group.bench_with_input(BenchmarkId::new("prove", len), &len, |b, &_| {
+            b.iter(|| {
+                let prover = create_proof_circuit::<
+                    KZGCommitmentScheme<_>,
+                    NLCircuit,
+                    ProverSHPLONK<_>,
+                    VerifierSHPLONK<_>,
+                    SingleStrategy<_>,
+                    _,
+                    EvmTranscript<_, _, _, _>,
+                    EvmTranscript<_, _, _, _>,
+                >(
+                    circuit.clone(),
+                    vec![],
+                    &params,
+                    &pk,
+                    CheckMode::UNSAFE,
+                    ezkl::Commitments::KZG,
+                    TranscriptType::EVM,
+                    None,
+                    None,
+                );
+                prover.unwrap();
+            });
+        });
+    }
+    group.finish();
+}
+
+criterion_group! {
+  name = benches;
+  config = Criterion::default().with_plots();
+  targets = runrelu
+}
+criterion_main!(benches);

benches/sigmoid.rs

@@ -2,6 +2,7 @@ use criterion::{criterion_group, criterion_main, BenchmarkId, Criterion, Through
 use ezkl::circuit::region::RegionCtx;
 use ezkl::circuit::table::Range;
 use ezkl::circuit::{ops::lookup::LookupOp, BaseConfig as Config, CheckMode};
+use ezkl::fieldutils::IntegerRep;
 use ezkl::pfsys::create_proof_circuit;
 use ezkl::pfsys::TranscriptType;
 use ezkl::pfsys::{create_keys, srs::gen_srs};
@@ -41,7 +42,7 @@ impl Circuit<Fr> for NLCircuit {
                 .map(|_| VarTensor::new_advice(cs, K, 1, LEN))
                 .collect::<Vec<_>>();
 
-            let nl = LookupOp::ReLU;
+            let nl = LookupOp::Sigmoid { scale: 1.0.into() };
 
             let mut config = Config::default();
 
@@ -62,9 +63,13 @@ impl Circuit<Fr> for NLCircuit {
         layouter.assign_region(
             || "",
             |region| {
-                let mut region = RegionCtx::new(region, 0, 1);
+                let mut region = RegionCtx::new(region, 0, 1, 1024, 2);
                 config
-                    .layout(&mut region, &[self.input.clone()], Box::new(LookupOp::ReLU))
+                    .layout(
+                        &mut region,
+                        &[self.input.clone()],
+                        Box::new(LookupOp::Sigmoid { scale: 1.0.into() }),
+                    )
                     .unwrap();
                 Ok(())
             },
@@ -84,7 +89,7 @@ fn runrelu(c: &mut Criterion) {
         };
 
         let input: Tensor<Value<Fr>> =
-            Tensor::<i32>::from((0..len).map(|_| rng.gen_range(0..10))).into();
+            Tensor::<IntegerRep>::from((0..len).map(|_| rng.gen_range(0..10))).into();
 
         let circuit = NLCircuit {
             input: ValTensor::from(input.clone()),

benches/zero_finder.rs

@@ -0,0 +1,117 @@
+use std::thread;
+
+use criterion::{black_box, criterion_group, criterion_main, Criterion};
+use halo2curves::{bn256::Fr as F, ff::Field};
+use maybe_rayon::{
+    iter::{IndexedParallelIterator, IntoParallelRefIterator, ParallelIterator},
+    slice::ParallelSlice,
+};
+use rand::Rng;
+
+// Assuming these are your types
+#[derive(Clone)]
+#[allow(dead_code)]
+enum ValType {
+    Constant(F),
+    AssignedConstant(usize, F),
+    Other,
+}
+
+// Helper to generate test data
+fn generate_test_data(size: usize, zero_probability: f64) -> Vec<ValType> {
+    let mut rng = rand::thread_rng();
+    (0..size)
+        .map(|_i| {
+            if rng.gen::<f64>() < zero_probability {
+                ValType::Constant(F::ZERO)
+            } else {
+                ValType::Constant(F::ONE) // Or some other non-zero value
+            }
+        })
+        .collect()
+}
+
+fn bench_zero_finding(c: &mut Criterion) {
+    let sizes = [
+        1_000,         // 1K
+        10_000,        // 10K
+        100_000,       // 100K
+        256 * 256 * 2, // Our specific case
+        1_000_000,     // 1M
+        10_000_000,    // 10M
+    ];
+
+    let zero_probability = 0.1; // 10% zeros
+
+    let mut group = c.benchmark_group("zero_finding");
+    group.sample_size(10); // Adjust based on your needs
+
+    for &size in &sizes {
+        let data = generate_test_data(size, zero_probability);
+
+        // Benchmark sequential version
+        group.bench_function(format!("sequential_{}", size), |b| {
+            b.iter(|| {
+                let result = data
+                    .iter()
+                    .enumerate()
+                    .filter_map(|(i, e)| match e {
+                        ValType::Constant(r) | ValType::AssignedConstant(_, r) => {
+                            (*r == F::ZERO).then_some(i)
+                        }
+                        _ => None,
+                    })
+                    .collect::<Vec<_>>();
+                black_box(result)
+            })
+        });
+
+        // Benchmark parallel version
+        group.bench_function(format!("parallel_{}", size), |b| {
+            b.iter(|| {
+                let result = data
+                    .par_iter()
+                    .enumerate()
+                    .filter_map(|(i, e)| match e {
+                        ValType::Constant(r) | ValType::AssignedConstant(_, r) => {
+                            (*r == F::ZERO).then_some(i)
+                        }
+                        _ => None,
+                    })
+                    .collect::<Vec<_>>();
+                black_box(result)
+            })
+        });
+
+        // Benchmark chunked parallel version
+        group.bench_function(format!("chunked_parallel_{}", size), |b| {
+            b.iter(|| {
+                let num_cores = thread::available_parallelism()
+                    .map(|n| n.get())
+                    .unwrap_or(1);
+                let chunk_size = (size / num_cores).max(100);
+
+                let result = data
+                    .par_chunks(chunk_size)
+                    .enumerate()
+                    .flat_map(|(chunk_idx, chunk)| {
+                        chunk
+                            .par_iter() // Make sure we use par_iter() here
+                            .enumerate()
+                            .filter_map(move |(i, e)| match e {
+                                ValType::Constant(r) | ValType::AssignedConstant(_, r) => {
+                                    (*r == F::ZERO).then_some(chunk_idx * chunk_size + i)
+                                }
+                                _ => None,
+                            })
+                    })
+                    .collect::<Vec<_>>();
+                black_box(result)
+            })
+        });
+    }
+    group.finish();
+}
+
+criterion_group!(benches, bench_zero_finding);
+criterion_main!(benches);

build.rs

@@ -0,0 +1,7 @@
+fn main() {
+    if cfg!(feature = "ios-bindings-test") {
+        println!("cargo::rustc-env=UNIFFI_CARGO_BUILD_EXTRA_ARGS=--features=ios-bindings --no-default-features");
+    }
+
+    println!("cargo::rerun-if-changed=build.rs");
+}

contracts/AttestData.sol

@@ -163,6 +163,253 @@ contract SwapProofCommitments {
     } /// end checkKzgCommits
 }
 
+contract DataAttestationSingle is LoadInstances, SwapProofCommitments {
+    /**
+     * @notice Struct used to make view only call to account to fetch the data that EZKL reads from.
+     * @param the address of the account to make calls to
+     * @param the abi encoded function calls to make to the `contractAddress`
+     */
+    struct AccountCall {
+        address contractAddress;
+        bytes callData;
+        uint256 decimals;
+    }
+    AccountCall public accountCall;
+
+    uint[] scales;
+
+    address public admin;
+
+    /**
+     * @notice EZKL P value
+     * @dev In order to prevent the verifier from accepting two version of the same pubInput, n and the quantity (n + P),  where n + P <= 2^256, we require that all instances are stricly less than P. a
+     * @dev The reason for this is that the assmebly code of the verifier performs all arithmetic operations modulo P and as a consequence can't distinguish between n and n + P.
+     */
+    uint256 constant ORDER =
+        uint256(
+            0x30644e72e131a029b85045b68181585d2833e84879b9709143e1f593f0000001
+        );
+
+    uint256 constant INPUT_LEN = 0;
+
+    uint256 constant OUTPUT_LEN = 0;
+
+    uint8 public instanceOffset;
+
+    /**
+     * @dev Initialize the contract with account calls the EZKL model will read from.
+     * @param _contractAddresses - The calls to all the contracts EZKL reads storage from.
+     * @param _callData - The abi encoded function calls to make to the `contractAddress` that EZKL reads storage from.
+     */
+    constructor(
+        address _contractAddresses,
+        bytes memory _callData,
+        uint256 _decimals,
+        uint[] memory _scales,
+        uint8 _instanceOffset,
+        address _admin
+    ) {
+        admin = _admin;
+        for (uint i; i < _scales.length; i++) {
+            scales.push(1 << _scales[i]);
+        }
+        populateAccountCalls(_contractAddresses, _callData, _decimals);
+        instanceOffset = _instanceOffset;
+    }
+
+    function updateAdmin(address _admin) external {
+        require(msg.sender == admin, "Only admin can update admin");
+        if (_admin == address(0)) {
+            revert();
+        }
+        admin = _admin;
+    }
+
+    function updateAccountCalls(
+        address _contractAddresses,
+        bytes memory _callData,
+        uint256 _decimals
+    ) external {
+        require(msg.sender == admin, "Only admin can update account calls");
+        populateAccountCalls(_contractAddresses, _callData, _decimals);
+    }
+
+    function populateAccountCalls(
+        address _contractAddresses,
+        bytes memory _callData,
+        uint256 _decimals
+    ) internal {
+        AccountCall memory _accountCall = accountCall;
+        _accountCall.contractAddress = _contractAddresses;
+        _accountCall.callData = _callData;
+        _accountCall.decimals = 10 ** _decimals;
+        accountCall = _accountCall;
+    }
+
+    function mulDiv(
+        uint256 x,
+        uint256 y,
+        uint256 denominator
+    ) internal pure returns (uint256 result) {
+        unchecked {
+            uint256 prod0;
+            uint256 prod1;
+            assembly {
+                let mm := mulmod(x, y, not(0))
+                prod0 := mul(x, y)
+                prod1 := sub(sub(mm, prod0), lt(mm, prod0))
+            }
+
+            if (prod1 == 0) {
+                return prod0 / denominator;
+            }
+
+            require(denominator > prod1, "Math: mulDiv overflow");
+
+            uint256 remainder;
+            assembly {
+                remainder := mulmod(x, y, denominator)
+                prod1 := sub(prod1, gt(remainder, prod0))
+                prod0 := sub(prod0, remainder)
+            }
+
+            uint256 twos = denominator & (~denominator + 1);
+            assembly {
+                denominator := div(denominator, twos)
+                prod0 := div(prod0, twos)
+                twos := add(div(sub(0, twos), twos), 1)
+            }
+
+            prod0 |= prod1 * twos;
+
+            uint256 inverse = (3 * denominator) ^ 2;
+
+            inverse *= 2 - denominator * inverse;
+            inverse *= 2 - denominator * inverse;
+            inverse *= 2 - denominator * inverse;
+            inverse *= 2 - denominator * inverse;
+            inverse *= 2 - denominator * inverse;
+            inverse *= 2 - denominator * inverse;
+
+            result = prod0 * inverse;
+            return result;
+        }
+    }
+    /**
+     * @dev Quantize the data returned from the account calls to the scale used by the EZKL model.
+     * @param x - One of the elements of the data returned from the account calls
+     * @param _decimals - Number of base 10 decimals to scale the data by.
+     * @param _scale - The base 2 scale used to convert the floating point value into a fixed point value.
+     *
+     */
+    function quantizeData(
+        int x,
+        uint256 _decimals,
+        uint256 _scale
+    ) internal pure returns (int256 quantized_data) {
+        bool neg = x < 0;
+        if (neg) x = -x;
+        uint output = mulDiv(uint256(x), _scale, _decimals);
+        if (mulmod(uint256(x), _scale, _decimals) * 2 >= _decimals) {
+            output += 1;
+        }
+        quantized_data = neg ? -int256(output) : int256(output);
+    }
+    /**
+     * @dev Make a static call to the account to fetch the data that EZKL reads from.
+     * @param target - The address of the account to make calls to.
+     * @param data  - The abi encoded function calls to make to the `contractAddress` that EZKL reads storage from.
+     * @return The data returned from the account calls. (Must come from either a view or pure function. Will throw an error otherwise)
+     */
+    function staticCall(
+        address target,
+        bytes memory data
+    ) internal view returns (bytes memory) {
+        (bool success, bytes memory returndata) = target.staticcall(data);
+        if (success) {
+            if (returndata.length == 0) {
+                require(
+                    target.code.length > 0,
+                    "Address: call to non-contract"
+                );
+            }
+            return returndata;
+        } else {
+            revert("Address: low-level call failed");
+        }
+    }
+    /**
+     * @dev Convert the fixed point quantized data into a field element.
+     * @param x - The quantized data.
+     * @return field_element - The field element.
+     */
+    function toFieldElement(
+        int256 x
+    ) internal pure returns (uint256 field_element) {
+        // The casting down to uint256 is safe because the order is about 2^254, and the value
+        // of x ranges of -2^127 to 2^127, so x + int(ORDER) is always positive.
+        return uint256(x + int(ORDER)) % ORDER;
+    }
+
+    /**
+     * @dev Make the account calls to fetch the data that EZKL reads from and attest to the data.
+     * @param instances - The public instances to the proof (the data in the proof that publicly accessible to the verifier).
+     */
+    function attestData(uint256[] memory instances) internal view {
+        require(
+            instances.length >= INPUT_LEN + OUTPUT_LEN,
+            "Invalid public inputs length"
+        );
+        AccountCall memory _accountCall = accountCall;
+        uint[] memory _scales = scales;
+        bytes memory returnData = staticCall(
+            _accountCall.contractAddress,
+            _accountCall.callData
+        );
+        int256[] memory x = abi.decode(returnData, (int256[]));
+        uint _offset;
+        int output = quantizeData(x[0], _accountCall.decimals, _scales[0]);
+        uint field_element = toFieldElement(output);
+        for (uint i = 0; i < x.length; i++) {
+            if (field_element != instances[i + instanceOffset]) {
+                _offset += 1;
+            } else {
+                break;
+            }
+        }
+        uint length = x.length - _offset;
+        for (uint i = 1; i < length; i++) {
+            output = quantizeData(x[i], _accountCall.decimals, _scales[i]);
+            field_element = toFieldElement(output);
+            require(
+                field_element == instances[i + instanceOffset + _offset],
+                "Public input does not match"
+            );
+        }
+    }
+
+    /**
+     * @dev Verify the proof with the data attestation.
+     * @param verifier - The address of the verifier contract.
+     * @param encoded - The verifier calldata.
+     */
+    function verifyWithDataAttestation(
+        address verifier,
+        bytes calldata encoded
+    ) public view returns (bool) {
+        require(verifier.code.length > 0, "Address: call to non-contract");
+        attestData(getInstancesCalldata(encoded));
+        // static call the verifier contract to verify the proof
+        (bool success, bytes memory returndata) = verifier.staticcall(encoded);
+
+        if (success) {
+            return abi.decode(returndata, (bool));
+        } else {
+            revert("low-level call to verifier failed");
+        }
+    }
+}
+
 // This contract serves as a Data Attestation Verifier for the EZKL model.
 // It is designed to read and attest to instances of proofs generated from a specified circuit.
 // It is particularly constructed to read only int256 data from specified on-chain contracts' view functions.
@@ -173,11 +420,11 @@ contract SwapProofCommitments {
 // 3. Static Calls: Makes static calls to fetch data from other contracts. See the `staticCall` method.
 // 4. Field Element Conversion: The fixed-point representation is then converted into a field element modulo P using the `toFieldElement` method.
 // 5. Data Attestation: The `attestData` method validates that the public instances match the data fetched and processed by the contract.
-// 6. Proof Verification: The `verifyWithDataAttestation` method parses the instances out of the encoded calldata and calls the `attestData` method to validate the public instances,
+// 6. Proof Verification: The `verifyWithDataAttestationMulti` method parses the instances out of the encoded calldata and calls the `attestData` method to validate the public instances,
 // 6b. Optional KZG Commitment Verification: It also checks the KZG commitments in the proof against the expected commitments using the `checkKzgCommits` method.
 //  then calls the `verifyProof` method to verify the proof on the verifier.
 
-contract DataAttestation is LoadInstances, SwapProofCommitments {
+contract DataAttestationMulti is LoadInstances, SwapProofCommitments {
     /**
      * @notice Struct used to make view only calls to accounts to fetch the data that EZKL reads from.
      * @param the address of the account to make calls to

docs/advanced_security/public_commitments.md

@@ -0,0 +1,41 @@
+## EZKL Security Note: Public Commitments and Low-Entropy Data
+
+> **Disclaimer:** this a more technical post that requires some prior knowledge of how ZK proving systems like Halo2 operate, and in particular in how these APIs are constructed. For background reading we highly recommend the [Halo2 book](https://zcash.github.io/halo2/) and [Halo2 Club](https://halo2.club/).
+
+## Overview of commitments in EZKL
+
+A common design pattern in a zero knowledge (zk) application is thus:
+- A prover has some data which is used within a circuit.
+- This data, as it may be high-dimensional or somewhat private, is pre-committed to using some hash function.
+- The zk-circuit which forms the core of the application then proves (para-phrasing) a statement of the form:
+>"I know some data D which when hashed corresponds to the pre-committed to value H + whatever else the circuit is proving over D". 
+
+From our own experience, we've implemented such patterns using snark-friendly hash functions like [Poseidon](https://www.poseidon-hash.info/), for which there is a relatively well vetted [implementation](https://docs.rs/halo2_gadgets/latest/halo2_gadgets/poseidon/index.html) in Halo2. Even then these hash functions can introduce lots of overhead and can be very expensive to generate proofs for if the dimensionality of the data D is large. 
+
+You can also implement such a pattern using Halo2's `Fixed` columns _if the privacy preservation of the pre-image is not necessary_. These are Halo2 columns (i.e in reality just polynomials) that are left unblinded (unlike the blinded `Advice` columns), and whose commitments are shared with the verifier by way of the verifying key for the application's zk-circuit. These commitments are much lower cost to generate than implementing a hashing function, such as Poseidon, within a circuit.
+
+> **Note:** Blinding is the process whereby a certain set of the final elements (i.e rows) of a Halo2 column are set to random field elements. This is the mechanism by which Halo2 achieves its zero knowledge properties for `Advice` columns. By contrast `Fixed` columns aren't zero-knowledge in that they are vulnerable to dictionary attacks in the same manner a hash function is. Given some set of known or popular data D an attacker can attempt to recover the pre-image of a hash by running D through the hash function to see if the outputs match a public commitment. These attacks aren't "possible" on blinded `Advice` columns.
+
+> **Further Note:** Note that without blinding, with access to `M` proofs, each of which contains an evaluation of the polynomial at a different point, an attacker can more easily recover a non blinded column's pre-image. This is because each proof generates a new query and evaluation of the polynomial represented by the column and as such with repetition a clearer picture can emerge of the column's pre-image. Thus unblinded columns should only be used for privacy preservation, in the manner of a hash, if the number of proofs generated against a fixed set of values is limited. More formally if M independent and _unique_ queries are generated; if M is equal to the degree + 1 of the polynomial represented by the column (i.e the unique lagrange interpolation of the values in the columns), then the column's pre-image can be recovered. As such as the logrows K increases, the more queries are required to recover the pre-image (as 2^K unique queries are required). This assumes that the entries in the column are not structured, as if they are then the number of queries required to recover the pre-image is reduced (eg. if all rows above a certain point are known to be nil).
+
+The annoyance in using `Fixed` columns comes from the fact that they require generating a new verifying key every time a new set of commitments is generated.
+
+> **Example:** Say for instance an application leverages a zero-knowledge circuit to prove the correct execution of a neural network. Every week the neural network is finetuned or retrained on new data. If the architecture remains the same then commiting to the new network parameters, along with a new proof of performance on a test set, would be an ideal setup. If we leverage `Fixed` columns to commit to the model parameters, each new commitment will require re-generating a  verifying key and sharing the new key with the verifier(s). This is not-ideal UX and can become expensive if the verifier is deployed on-chain. 
+
+An ideal commitment would thus have the low cost of a `Fixed`  column but wouldn't require regenerating a new verifying key for each new commitment.
+
+### Unblinded Advice Columns
+
+A first step in designing such a commitment is to allow for optionally unblinded `Advice` columns within the Halo2 API. These won't be included in the verifying key, AND are blinded with a constant factor `1` -- such that if someone knows the pre-image to the commitment, they can recover it by running it through the corresponding polynomial commitment scheme (in ezkl's case [KZG commitments](https://dankradfeist.de/ethereum/2020/06/16/kate-polynomial-commitments.html)). 
+
+This is implemented using the `polycommit` visibility parameter in the ezkl API.
+
+## The Vulnerability of Public Commitments
+
+
+Public commitments in EZKL (both Poseidon-hashed inputs and KZG commitments) can be vulnerable to brute-force attacks when input data has low entropy. A malicious actor could reveal committed data by searching through possible input values, compromising privacy in applications like anonymous credentials. This is particularly relevant when input data comes from known finite sets (e.g., names, dates).
+
+Example Risk: In an anonymous credential system using EZKL for ID verification, an attacker could match hashed outputs against a database of common identifying information to deanonymize users.
+
+
+

docs/advanced_security/quantization_backdoors.md

@@ -0,0 +1,22 @@
+# EZKL Security Note: Quantization-Induced Model Backdoors
+
+> Note: this only affects a situation where a party separate to an application's developer has access to the model's weights and can modify them. This is a common scenario in adversarial machine learning research, but can be less common in real-world applications. If you're building your models in house and deploying them yourself, this is less of a concern. If you're building a permisionless system where anyone can submit models, this is more of a concern.
+
+Models processed through EZKL's quantization step can harbor backdoors that are dormant in the original full-precision model but activate during quantization. These backdoors force specific outputs when triggered, with impact varying by application.
+
+Key Factors:
+
+- Larger models increase attack feasibility through more parameter capacity
+- Smaller quantization scales facilitate attacks by allowing greater weight modifications
+- Rebase ratio of 1 enables exploitation of convolutional layer consistency
+
+Limitations: 
+
+- Attack effectiveness depends on calibration settings and internal rescaling operations. 
+- Further research needed on backdoor persistence through witness/proof stages.
+- Can be mitigated by evaluating the quantized model (using `ezkl gen-witness`), rather than relying on the evaluation of the original model. 
+
+References:
+
+1. [Quantization Backdoors to Deep Learning Commercial Frameworks (Ma et al., 2021)](https://arxiv.org/abs/2108.09187)
+2. [Planting Undetectable Backdoors in Machine Learning Models (Goldwasser et al., 2022)](https://arxiv.org/abs/2204.06974)

docs/python/requirements-docs.txt

@@ -1,4 +1,4 @@
-ezkl==11.6.0
+ezkl
 sphinx
 sphinx-rtd-theme
 sphinxcontrib-napoleon

docs/python/src/conf.py

@@ -1,7 +1,7 @@
 import ezkl
 
 project = 'ezkl'
-release = '11.6.0'
+release = '0.0.0'
 version = release
 
 

examples/conv2d_mnist/main.rs

@@ -2,8 +2,7 @@ use ezkl::circuit::region::RegionCtx;
 use ezkl::circuit::{
     ops::lookup::LookupOp, ops::poly::PolyOp, BaseConfig as PolyConfig, CheckMode,
 };
-use ezkl::fieldutils;
-use ezkl::fieldutils::i32_to_felt;
+use ezkl::fieldutils::{self, integer_rep_to_felt, IntegerRep};
 use ezkl::tensor::*;
 use halo2_proofs::dev::MockProver;
 use halo2_proofs::poly::commitment::Params;
@@ -33,6 +32,7 @@ use mnist::*;
 use rand::rngs::OsRng;
 use std::marker::PhantomData;
 
+
 mod params;
 
 const K: usize = 20;
@@ -42,8 +42,8 @@ const NUM_INNER_COLS: usize = 1;
 struct Config<
     const LEN: usize, //LEN = CHOUT x OH x OW flattened //not supported yet in rust stable
     const CLASSES: usize,
-    const LOOKUP_MIN: i64,
-    const LOOKUP_MAX: i64,
+    const LOOKUP_MIN: IntegerRep,
+    const LOOKUP_MAX: IntegerRep,
     // Convolution
     const KERNEL_HEIGHT: usize,
     const KERNEL_WIDTH: usize,
@@ -66,8 +66,8 @@ struct Config<
 struct MyCircuit<
     const LEN: usize, //LEN = CHOUT x OH x OW flattened
     const CLASSES: usize,
-    const LOOKUP_MIN: i64,
-    const LOOKUP_MAX: i64,
+    const LOOKUP_MIN: IntegerRep,
+    const LOOKUP_MAX: IntegerRep,
     // Convolution
     const KERNEL_HEIGHT: usize,
     const KERNEL_WIDTH: usize,
@@ -90,8 +90,8 @@ struct MyCircuit<
 impl<
         const LEN: usize,
         const CLASSES: usize,
-        const LOOKUP_MIN: i64,
-        const LOOKUP_MAX: i64,
+        const LOOKUP_MIN: IntegerRep,
+        const LOOKUP_MAX: IntegerRep,
         // Convolution
         const KERNEL_HEIGHT: usize,
         const KERNEL_WIDTH: usize,
@@ -147,6 +147,8 @@ where
         let params = VarTensor::new_advice(cs, K, NUM_INNER_COLS, LEN);
         let output = VarTensor::new_advice(cs, K, NUM_INNER_COLS, LEN);
 
+        let _constant = VarTensor::constant_cols(cs, K, LEN, false);
+
         println!("INPUT COL {:#?}", input);
 
         let mut layer_config = PolyConfig::configure(
@@ -157,15 +159,11 @@ where
         );
 
         layer_config
-            .configure_lookup(
-                cs,
-                &input,
-                &output,
-                &params,
-                (LOOKUP_MIN, LOOKUP_MAX),
-                K,
-                &LookupOp::ReLU,
-            )
+            .configure_range_check(cs, &input, &params, (-1, 1), K)
+            .unwrap();
+
+        layer_config
+            .configure_range_check(cs, &input, &params, (0, 1023), K)
             .unwrap();
 
         layer_config
@@ -196,15 +194,23 @@ where
     ) -> Result<(), Error> {
         config.layer_config.layout_tables(&mut layouter).unwrap();
 
+        config
+            .layer_config
+            .layout_range_checks(&mut layouter)
+            .unwrap();
+
         let x = layouter
             .assign_region(
                 || "mlp_4d",
                 |region| {
-                    let mut region = RegionCtx::new(region, 0, NUM_INNER_COLS);
+                    let mut region = RegionCtx::new(region, 0, NUM_INNER_COLS, 1024, 2);
 
                     let op = PolyOp::Conv {
                         padding: vec![(PADDING, PADDING); 2],
                         stride: vec![STRIDE; 2],
+                        group: 1,
+                        data_format: DataFormat::NCHW,
+                        kernel_format: KernelFormat::OIHW,
                     };
                     let x = config
                         .layer_config
@@ -221,7 +227,14 @@ where
 
                     let x = config
                         .layer_config
-                        .layout(&mut region, &[x.unwrap()], Box::new(LookupOp::ReLU))
+                        .layout(
+                            &mut region,
+                            &[x.unwrap()],
+                            Box::new(PolyOp::LeakyReLU {
+                                slope: 0.0.into(),
+                                scale: 1,
+                            }),
+                        )
                         .unwrap();
 
                     let mut x = config
@@ -281,7 +294,7 @@ where
 }
 
 pub fn runconv() {
-    #[cfg(not(target_arch = "wasm32"))]
+    #[cfg(all(feature = "ezkl", not(target_arch = "wasm32")))]
     env_logger::init();
 
     const KERNEL_HEIGHT: usize = 5;
@@ -315,7 +328,11 @@ pub fn runconv() {
         .test_set_length(10_000)
         .finalize();
 
-    let mut train_data = Tensor::from(trn_img.iter().map(|x| i32_to_felt::<F>(*x as i32 / 16)));
+    let mut train_data = Tensor::from(
+        trn_img
+            .iter()
+            .map(|x| integer_rep_to_felt::<F>(*x as IntegerRep / 16)),
+    );
     train_data.reshape(&[50_000, 28, 28]).unwrap();
 
     let mut train_labels = Tensor::from(trn_lbl.iter().map(|x| *x as f32));
@@ -343,8 +360,8 @@ pub fn runconv() {
             .map(|fl| {
                 let dx = fl * 32_f32;
                 let rounded = dx.round();
-                let integral: i32 = unsafe { rounded.to_int_unchecked() };
-                fieldutils::i32_to_felt(integral)
+                let integral: IntegerRep = unsafe { rounded.to_int_unchecked() };
+                fieldutils::integer_rep_to_felt(integral)
             }),
     );
 
@@ -355,7 +372,8 @@ pub fn runconv() {
 
     let l0_kernels = l0_kernels.try_into().unwrap();
 
-    let mut l0_bias = Tensor::<F>::from((0..OUT_CHANNELS).map(|_| fieldutils::i32_to_felt(0)));
+    let mut l0_bias =
+        Tensor::<F>::from((0..OUT_CHANNELS).map(|_| fieldutils::integer_rep_to_felt(0)));
     l0_bias.set_visibility(&ezkl::graph::Visibility::Private);
 
     let l0_bias = l0_bias.try_into().unwrap();
@@ -363,8 +381,8 @@ pub fn runconv() {
     let mut l2_biases = Tensor::<F>::from(myparams.biases.into_iter().map(|fl| {
         let dx = fl * 32_f32;
         let rounded = dx.round();
-        let integral: i32 = unsafe { rounded.to_int_unchecked() };
-        fieldutils::i32_to_felt(integral)
+        let integral: IntegerRep = unsafe { rounded.to_int_unchecked() };
+        fieldutils::integer_rep_to_felt(integral)
     }));
     l2_biases.set_visibility(&ezkl::graph::Visibility::Private);
     l2_biases.reshape(&[l2_biases.len(), 1]).unwrap();
@@ -374,8 +392,8 @@ pub fn runconv() {
     let mut l2_weights = Tensor::<F>::from(myparams.weights.into_iter().flatten().map(|fl| {
         let dx = fl * 32_f32;
         let rounded = dx.round();
-        let integral: i32 = unsafe { rounded.to_int_unchecked() };
-        fieldutils::i32_to_felt(integral)
+        let integral: IntegerRep = unsafe { rounded.to_int_unchecked() };
+        fieldutils::integer_rep_to_felt(integral)
     }));
     l2_weights.set_visibility(&ezkl::graph::Visibility::Private);
     l2_weights.reshape(&[CLASSES, LEN]).unwrap();
@@ -401,13 +419,13 @@ pub fn runconv() {
         l2_params: [l2_weights, l2_biases],
     };
 
-    let public_input: Tensor<i32> = vec![
-        -25124i32, -19304, -16668, -4399, -6209, -4548, -2317, -8349, -6117, -23461,
+    let public_input: Tensor<IntegerRep> = vec![
+        -25124, -19304, -16668, -4399, -6209, -4548, -2317, -8349, -6117, -23461,
     ]
     .into_iter()
     .into();
 
-    let pi_inner: Tensor<F> = public_input.map(i32_to_felt::<F>);
+    let pi_inner: Tensor<F> = public_input.map(integer_rep_to_felt::<F>);
 
     println!("MOCK PROVING");
     let now = Instant::now();

examples/mlp_4d_einsum.rs

@@ -2,7 +2,7 @@ use ezkl::circuit::region::RegionCtx;
 use ezkl::circuit::{
     ops::lookup::LookupOp, ops::poly::PolyOp, BaseConfig as PolyConfig, CheckMode,
 };
-use ezkl::fieldutils::i32_to_felt;
+use ezkl::fieldutils::{integer_rep_to_felt, IntegerRep};
 use ezkl::tensor::*;
 use halo2_proofs::dev::MockProver;
 use halo2_proofs::{
@@ -23,8 +23,8 @@ struct MyConfig {
 #[derive(Clone)]
 struct MyCircuit<
     const LEN: usize, //LEN = CHOUT x OH x OW flattened
-    const LOOKUP_MIN: i64,
-    const LOOKUP_MAX: i64,
+    const LOOKUP_MIN: IntegerRep,
+    const LOOKUP_MAX: IntegerRep,
 > {
     // Given the stateless MyConfig type information, a DNN trace is determined by its input and the parameters of its layers.
     // Computing the trace still requires a forward pass. The intermediate activations are stored only by the layouter.
@@ -34,7 +34,7 @@ struct MyCircuit<
     _marker: PhantomData<F>,
 }
 
-impl<const LEN: usize, const LOOKUP_MIN: i64, const LOOKUP_MAX: i64> Circuit<F>
+impl<const LEN: usize, const LOOKUP_MIN: IntegerRep, const LOOKUP_MAX: IntegerRep> Circuit<F>
     for MyCircuit<LEN, LOOKUP_MIN, LOOKUP_MAX>
 {
     type Config = MyConfig;
@@ -53,6 +53,10 @@ impl<const LEN: usize, const LOOKUP_MIN: i64, const LOOKUP_MAX: i64> Circuit<F>
         let output = VarTensor::new_advice(cs, K, 1, LEN);
         // tells the config layer to add an affine op to the circuit gate
 
+        let _constant = VarTensor::constant_cols(cs, K, LEN, false);
+
+        println!("INPUT COL {:#?}", input);
+
         let mut layer_config = PolyConfig::<F>::configure(
             cs,
             &[input.clone(), params.clone()],
@@ -60,17 +64,12 @@ impl<const LEN: usize, const LOOKUP_MIN: i64, const LOOKUP_MAX: i64> Circuit<F>
             CheckMode::SAFE,
         );
 
-        // sets up a new ReLU table and resuses it for l1 and l3 non linearities
         layer_config
-            .configure_lookup(
-                cs,
-                &input,
-                &output,
-                &params,
-                (LOOKUP_MIN, LOOKUP_MAX),
-                K,
-                &LookupOp::ReLU,
-            )
+            .configure_range_check(cs, &input, &params, (-1, 1), K)
+            .unwrap();
+
+        layer_config
+            .configure_range_check(cs, &input, &params, (0, 1023), K)
             .unwrap();
 
         // sets up a new ReLU table and resuses it for l1 and l3 non linearities
@@ -104,11 +103,16 @@ impl<const LEN: usize, const LOOKUP_MIN: i64, const LOOKUP_MAX: i64> Circuit<F>
     ) -> Result<(), Error> {
         config.layer_config.layout_tables(&mut layouter).unwrap();
 
+        config
+            .layer_config
+            .layout_range_checks(&mut layouter)
+            .unwrap();
+
         let x = layouter
             .assign_region(
                 || "mlp_4d",
                 |region| {
-                    let mut region = RegionCtx::new(region, 0, 1);
+                    let mut region = RegionCtx::new(region, 0, 1, 1024, 2);
                     let x = config
                         .layer_config
                         .layout(
@@ -141,7 +145,14 @@ impl<const LEN: usize, const LOOKUP_MIN: i64, const LOOKUP_MAX: i64> Circuit<F>
                     println!("x shape: {:?}", x.dims());
                     let mut x = config
                         .layer_config
-                        .layout(&mut region, &[x], Box::new(LookupOp::ReLU))
+                        .layout(
+                            &mut region,
+                            &[x],
+                            Box::new(PolyOp::LeakyReLU {
+                                scale: 1,
+                                slope: 0.0.into(),
+                            }),
+                        )
                         .unwrap()
                         .unwrap();
                     println!("3");
@@ -177,7 +188,14 @@ impl<const LEN: usize, const LOOKUP_MIN: i64, const LOOKUP_MAX: i64> Circuit<F>
                     println!("x shape: {:?}", x.dims());
                     let x = config
                         .layer_config
-                        .layout(&mut region, &[x], Box::new(LookupOp::ReLU))
+                        .layout(
+                            &mut region,
+                            &[x],
+                            Box::new(PolyOp::LeakyReLU {
+                                scale: 1,
+                                slope: 0.0.into(),
+                            }),
+                        )
                         .unwrap();
                     println!("6");
                     println!("offset: {}", region.row());
@@ -212,36 +230,36 @@ impl<const LEN: usize, const LOOKUP_MIN: i64, const LOOKUP_MAX: i64> Circuit<F>
 }
 
 pub fn runmlp() {
-    #[cfg(not(target_arch = "wasm32"))]
+    #[cfg(all(feature = "ezkl", not(target_arch = "wasm32")))]
     env_logger::init();
     // parameters
-    let mut l0_kernel: Tensor<F> = Tensor::<i32>::new(
+    let mut l0_kernel: Tensor<F> = Tensor::<IntegerRep>::new(
         Some(&[10, 0, 0, -1, 0, 10, 1, 0, 0, 1, 10, 0, 1, 0, 0, 10]),
         &[4, 4],
     )
     .unwrap()
-    .map(i32_to_felt);
+    .map(integer_rep_to_felt);
     l0_kernel.set_visibility(&ezkl::graph::Visibility::Private);
 
-    let mut l0_bias: Tensor<F> = Tensor::<i32>::new(Some(&[0, 0, 0, 1]), &[4, 1])
+    let mut l0_bias: Tensor<F> = Tensor::<IntegerRep>::new(Some(&[0, 0, 0, 1]), &[4, 1])
         .unwrap()
-        .map(i32_to_felt);
+        .map(integer_rep_to_felt);
     l0_bias.set_visibility(&ezkl::graph::Visibility::Private);
 
-    let mut l2_kernel: Tensor<F> = Tensor::<i32>::new(
+    let mut l2_kernel: Tensor<F> = Tensor::<IntegerRep>::new(
         Some(&[0, 3, 10, -1, 0, 10, 1, 0, 0, 1, 0, 12, 1, -2, 32, 0]),
         &[4, 4],
     )
     .unwrap()
-    .map(i32_to_felt);
+    .map(integer_rep_to_felt);
     l2_kernel.set_visibility(&ezkl::graph::Visibility::Private);
     // input data, with 1 padding to allow for bias
-    let input: Tensor<Value<F>> = Tensor::<i32>::new(Some(&[-30, -21, 11, 40]), &[4, 1])
+    let input: Tensor<Value<F>> = Tensor::<IntegerRep>::new(Some(&[-30, -21, 11, 40]), &[4, 1])
         .unwrap()
         .into();
-    let mut l2_bias: Tensor<F> = Tensor::<i32>::new(Some(&[0, 0, 0, 1]), &[4, 1])
+    let mut l2_bias: Tensor<F> = Tensor::<IntegerRep>::new(Some(&[0, 0, 0, 1]), &[4, 1])
         .unwrap()
-        .map(i32_to_felt);
+        .map(integer_rep_to_felt);
     l2_bias.set_visibility(&ezkl::graph::Visibility::Private);
 
     let circuit = MyCircuit::<4, -8192, 8192> {
@@ -251,12 +269,12 @@ pub fn runmlp() {
         _marker: PhantomData,
     };
 
-    let public_input: Vec<i32> = unsafe {
+    let public_input: Vec<IntegerRep> = unsafe {
         vec![
-            (531f32 / 128f32).round().to_int_unchecked::<i32>(),
-            (103f32 / 128f32).round().to_int_unchecked::<i32>(),
-            (4469f32 / 128f32).round().to_int_unchecked::<i32>(),
-            (2849f32 / 128f32).to_int_unchecked::<i32>(),
+            (531f32 / 128f32).round().to_int_unchecked::<IntegerRep>(),
+            (103f32 / 128f32).round().to_int_unchecked::<IntegerRep>(),
+            (4469f32 / 128f32).round().to_int_unchecked::<IntegerRep>(),
+            (2849f32 / 128f32).to_int_unchecked::<IntegerRep>(),
         ]
     };
 
@@ -265,7 +283,10 @@ pub fn runmlp() {
     let prover = MockProver::run(
         K as u32,
         &circuit,
-        vec![public_input.iter().map(|x| i32_to_felt::<F>(*x)).collect()],
+        vec![public_input
+            .iter()
+            .map(|x| integer_rep_to_felt::<F>(*x))
+            .collect()],
     )
     .unwrap();
     prover.assert_satisfied();

examples/notebooks/cat_and_dog_data.sh

@@ -0,0 +1,13 @@
+# download tess data
+# check if first argument has been set 
+if [ ! -z "$1" ]; then
+    DATA_DIR=$1
+else
+    DATA_DIR=data
+fi
+
+echo "Downloading data to $DATA_DIR"
+
+if [ ! -d "$DATA_DIR/CATDOG" ]; then
+    kaggle datasets download tongpython/cat-and-dog -p $DATA_DIR/CATDOG --unzip
+fi

examples/notebooks/data_attest.ipynb

@@ -592,7 +592,7 @@
             "name": "python",
             "nbconvert_exporter": "python",
             "pygments_lexer": "ipython3",
-            "version": "3.12.2"
+            "version": "3.12.5"
         },
         "orig_nbformat": 4
     },

examples/notebooks/data_attest_hashed.ipynb

@@ -648,10 +648,10 @@
             "name": "python",
             "nbconvert_exporter": "python",
             "pygments_lexer": "ipython3",
-            "version": "3.9.15"
+            "version": "3.12.7"
         },
         "orig_nbformat": 4
     },
     "nbformat": 4,
     "nbformat_minor": 2
-}
+}

examples/notebooks/generalized_inverse.ipynb

@@ -77,6 +77,7 @@
       "outputs": [],
       "source": [
         "gip_run_args = ezkl.PyRunArgs()\n",
+        "gip_run_args.ignore_range_check_inputs_outputs = True\n",
         "gip_run_args.input_visibility = \"polycommit\"  # matrix and generalized inverse commitments\n",
         "gip_run_args.output_visibility = \"fixed\"   # no parameters used\n",
         "gip_run_args.param_visibility = \"fixed\" # should be Tensor(True)"
@@ -335,9 +336,9 @@
       "name": "python",
       "nbconvert_exporter": "python",
       "pygments_lexer": "ipython3",
-      "version": "3.9.15"
+      "version": "3.9.13"
     }
   },
   "nbformat": 4,
   "nbformat_minor": 5
-}
+}

examples/notebooks/linear_regression.ipynb

@@ -1,279 +1,284 @@
 {
-    "cells": [
-        {
-            "cell_type": "markdown",
-            "id": "cf69bb3f-94e6-4dba-92cd-ce08df117d67",
-            "metadata": {},
-            "source": [
-                "## Linear Regression\n",
-                "\n",
-                "\n",
-                "Sklearn based models are slightly finicky to get into a suitable onnx format. \n",
-                "This notebook showcases how to do so using the `hummingbird-ml` python package ! "
-            ]
-        },
-        {
-            "cell_type": "code",
-            "execution_count": null,
-            "id": "95613ee9",
-            "metadata": {},
-            "outputs": [],
-            "source": [
-                "# check if notebook is in colab\n",
-                "try:\n",
-                "    # install ezkl\n",
-                "    import google.colab\n",
-                "    import subprocess\n",
-                "    import sys\n",
-                "    subprocess.check_call([sys.executable, \"-m\", \"pip\", \"install\", \"ezkl\"])\n",
-                "    subprocess.check_call([sys.executable, \"-m\", \"pip\", \"install\", \"onnx\"])\n",
-                "    subprocess.check_call([sys.executable, \"-m\", \"pip\", \"install\", \"hummingbird-ml\"])\n",
-                "\n",
-                "# rely on local installation of ezkl if the notebook is not in colab\n",
-                "except:\n",
-                "    pass\n",
-                "\n",
-                "import os\n",
-                "import torch\n",
-                "import ezkl\n",
-                "import json\n",
-                "from hummingbird.ml import convert\n",
-                "\n",
-                "\n",
-                "# here we create and (potentially train a model)\n",
-                "\n",
-                "# make sure you have the dependencies required here already installed\n",
-                "import numpy as np\n",
-                "from sklearn.linear_model import LinearRegression\n",
-                "X = np.array([[1, 1], [1, 2], [2, 2], [2, 3]])\n",
-                "# y = 1 * x_0 + 2 * x_1 + 3\n",
-                "y = np.dot(X, np.array([1, 2])) + 3\n",
-                "reg = LinearRegression().fit(X, y)\n",
-                "reg.score(X, y)\n",
-                "\n",
-                "circuit = convert(reg, \"torch\", X[:1]).model\n",
-                "\n",
-                "\n",
-                "\n"
-            ]
-        },
-        {
-            "cell_type": "code",
-            "execution_count": null,
-            "id": "b37637c4",
-            "metadata": {},
-            "outputs": [],
-            "source": [
-                "model_path = os.path.join('network.onnx')\n",
-                "compiled_model_path = os.path.join('network.compiled')\n",
-                "pk_path = os.path.join('test.pk')\n",
-                "vk_path = os.path.join('test.vk')\n",
-                "settings_path = os.path.join('settings.json')\n",
-                "\n",
-                "witness_path = os.path.join('witness.json')\n",
-                "data_path = os.path.join('input.json')"
-            ]
-        },
-        {
-            "cell_type": "code",
-            "execution_count": null,
-            "id": "82db373a",
-            "metadata": {},
-            "outputs": [],
-            "source": [
-                "\n",
-                "\n",
-                "# export to onnx format\n",
-                "# !!!!!!!!!!!!!!!!! This will flash a warning but it is fine !!!!!!!!!!!!!!!!!!!!!\n",
-                "\n",
-                "# Input to the model\n",
-                "shape = X.shape[1:]\n",
-                "x = torch.rand(1, *shape, requires_grad=True)\n",
-                "torch_out = circuit(x)\n",
-                "# Export the model\n",
-                "torch.onnx.export(circuit,               # model being run\n",
-                "                  # model input (or a tuple for multiple inputs)\n",
-                "                  x,\n",
-                "                  # where to save the model (can be a file or file-like object)\n",
-                "                  \"network.onnx\",\n",
-                "                  export_params=True,        # store the trained parameter weights inside the model file\n",
-                "                  opset_version=10,          # the ONNX version to export the model to\n",
-                "                  do_constant_folding=True,  # whether to execute constant folding for optimization\n",
-                "                  input_names=['input'],   # the model's input names\n",
-                "                  output_names=['output'],  # the model's output names\n",
-                "                  dynamic_axes={'input': {0: 'batch_size'},    # variable length axes\n",
-                "                                'output': {0: 'batch_size'}})\n",
-                "\n",
-                "d = ((x).detach().numpy()).reshape([-1]).tolist()\n",
-                "\n",
-                "data = dict(input_shapes=[shape],\n",
-                "            input_data=[d],\n",
-                "            output_data=[((o).detach().numpy()).reshape([-1]).tolist() for o in torch_out])\n",
-                "\n",
-                "# Serialize data into file:\n",
-                "json.dump(data, open(\"input.json\", 'w'))\n"
-            ]
-        },
-        {
-            "cell_type": "code",
-            "execution_count": null,
-            "id": "d5e374a2",
-            "metadata": {},
-            "outputs": [],
-            "source": [
-                "!RUST_LOG=trace\n",
-                "# TODO: Dictionary outputs\n",
-                "res = ezkl.gen_settings(model_path, settings_path)\n",
-                "assert res == True\n"
-            ]
-        },
-        {
-            "cell_type": "code",
-            "execution_count": null,
-            "metadata": {},
-            "outputs": [],
-            "source": [
-                "cal_path = os.path.join(\"calibration.json\")\n",
-                "\n",
-                "data_array = (torch.randn(20, *shape).detach().numpy()).reshape([-1]).tolist()\n",
-                "\n",
-                "data = dict(input_data = [data_array])\n",
-                "\n",
-                "# Serialize data into file:\n",
-                "json.dump(data, open(cal_path, 'w'))\n",
-                "\n",
-                "res = await ezkl.calibrate_settings(data_path, model_path, settings_path, \"resources\")\n",
-                "assert res == True\n"
-            ]
-        },
-        {
-            "cell_type": "code",
-            "execution_count": null,
-            "id": "3aa4f090",
-            "metadata": {},
-            "outputs": [],
-            "source": [
-                "res = ezkl.compile_circuit(model_path, compiled_model_path, settings_path)\n",
-                "assert res == True"
-            ]
-        },
-        {
-            "cell_type": "code",
-            "execution_count": null,
-            "id": "8b74dcee",
-            "metadata": {},
-            "outputs": [],
-            "source": [
-                "# srs path\n",
-                "res = await ezkl.get_srs( settings_path)"
-            ]
-        },
-        {
-            "cell_type": "code",
-            "execution_count": null,
-            "id": "18c8b7c7",
-            "metadata": {},
-            "outputs": [],
-            "source": [
-                "# now generate the witness file \n",
-                "\n",
-                "res = await ezkl.gen_witness(data_path, compiled_model_path, witness_path)\n",
-                "assert os.path.isfile(witness_path)"
-            ]
-        },
-        {
-            "cell_type": "code",
-            "execution_count": null,
-            "id": "b1c561a8",
-            "metadata": {},
-            "outputs": [],
-            "source": [
-                "\n",
-                "# HERE WE SETUP THE CIRCUIT PARAMS\n",
-                "# WE GOT KEYS\n",
-                "# WE GOT CIRCUIT PARAMETERS\n",
-                "# EVERYTHING ANYONE HAS EVER NEEDED FOR ZK\n",
-                "\n",
-                "\n",
-                "\n",
-                "res = ezkl.setup(\n",
-                "        compiled_model_path,\n",
-                "        vk_path,\n",
-                "        pk_path,\n",
-                "        \n",
-                "    )\n",
-                "\n",
-                "assert res == True\n",
-                "assert os.path.isfile(vk_path)\n",
-                "assert os.path.isfile(pk_path)\n",
-                "assert os.path.isfile(settings_path)"
-            ]
-        },
-        {
-            "cell_type": "code",
-            "execution_count": null,
-            "id": "c384cbc8",
-            "metadata": {},
-            "outputs": [],
-            "source": [
-                "# GENERATE A PROOF\n",
-                "\n",
-                "\n",
-                "proof_path = os.path.join('test.pf')\n",
-                "\n",
-                "res = ezkl.prove(\n",
-                "        witness_path,\n",
-                "        compiled_model_path,\n",
-                "        pk_path,\n",
-                "        proof_path,\n",
-                "        \n",
-                "        \"single\",\n",
-                "    )\n",
-                "\n",
-                "print(res)\n",
-                "assert os.path.isfile(proof_path)"
-            ]
-        },
-        {
-            "cell_type": "code",
-            "execution_count": null,
-            "id": "76f00d41",
-            "metadata": {},
-            "outputs": [],
-            "source": [
-                "# VERIFY IT\n",
-                "\n",
-                "res = ezkl.verify(\n",
-                "        proof_path,\n",
-                "        settings_path,\n",
-                "        vk_path,\n",
-                "        \n",
-                "    )\n",
-                "\n",
-                "assert res == True\n",
-                "print(\"verified\")"
-            ]
-        }
-    ],
-    "metadata": {
-        "kernelspec": {
-            "display_name": "Python 3 (ipykernel)",
-            "language": "python",
-            "name": "python3"
-        },
-        "language_info": {
-            "codemirror_mode": {
-                "name": "ipython",
-                "version": 3
-            },
-            "file_extension": ".py",
-            "mimetype": "text/x-python",
-            "name": "python",
-            "nbconvert_exporter": "python",
-            "pygments_lexer": "ipython3",
-            "version": "3.9.15"
-        }
-    },
-    "nbformat": 4,
-    "nbformat_minor": 5
-}
+ "cells": [
+  {
+   "cell_type": "markdown",
+   "id": "cf69bb3f-94e6-4dba-92cd-ce08df117d67",
+   "metadata": {},
+   "source": [
+    "## Linear Regression\n",
+    "\n",
+    "\n",
+    "Sklearn based models are slightly finicky to get into a suitable onnx format. \n",
+    "This notebook showcases how to do so using the `hummingbird-ml` python package ! "
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "95613ee9",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "# check if notebook is in colab\n",
+    "try:\n",
+    "    # install ezkl\n",
+    "    import google.colab\n",
+    "    import subprocess\n",
+    "    import sys\n",
+    "    subprocess.check_call([sys.executable, \"-m\", \"pip\", \"install\", \"ezkl\"])\n",
+    "    subprocess.check_call([sys.executable, \"-m\", \"pip\", \"install\", \"onnx\"])\n",
+    "    subprocess.check_call([sys.executable, \"-m\", \"pip\", \"install\", \"hummingbird-ml\"])\n",
+    "\n",
+    "# rely on local installation of ezkl if the notebook is not in colab\n",
+    "except:\n",
+    "    pass\n",
+    "\n",
+    "import os\n",
+    "import torch\n",
+    "import ezkl\n",
+    "import json\n",
+    "from hummingbird.ml import convert\n",
+    "\n",
+    "\n",
+    "# here we create and (potentially train a model)\n",
+    "\n",
+    "# make sure you have the dependencies required here already installed\n",
+    "import numpy as np\n",
+    "from sklearn.linear_model import LinearRegression\n",
+    "X = np.array([[1, 1], [1, 2], [2, 2], [2, 3]])\n",
+    "# y = 1 * x_0 + 2 * x_1 + 3\n",
+    "y = np.dot(X, np.array([1, 2])) + 3\n",
+    "reg = LinearRegression().fit(X, y)\n",
+    "reg.score(X, y)\n",
+    "\n",
+    "circuit = convert(reg, \"torch\", X[:1]).model\n",
+    "\n",
+    "\n",
+    "\n"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "b37637c4",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "model_path = os.path.join('network.onnx')\n",
+    "compiled_model_path = os.path.join('network.compiled')\n",
+    "pk_path = os.path.join('test.pk')\n",
+    "vk_path = os.path.join('test.vk')\n",
+    "settings_path = os.path.join('settings.json')\n",
+    "\n",
+    "witness_path = os.path.join('witness.json')\n",
+    "data_path = os.path.join('input.json')"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "82db373a",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "\n",
+    "\n",
+    "# export to onnx format\n",
+    "# !!!!!!!!!!!!!!!!! This will flash a warning but it is fine !!!!!!!!!!!!!!!!!!!!!\n",
+    "\n",
+    "# Input to the model\n",
+    "shape = X.shape[1:]\n",
+    "x = torch.rand(1, *shape, requires_grad=True)\n",
+    "torch_out = circuit(x)\n",
+    "# Export the model\n",
+    "torch.onnx.export(circuit,               # model being run\n",
+    "                  # model input (or a tuple for multiple inputs)\n",
+    "                  x,\n",
+    "                  # where to save the model (can be a file or file-like object)\n",
+    "                  \"network.onnx\",\n",
+    "                  export_params=True,        # store the trained parameter weights inside the model file\n",
+    "                  opset_version=10,          # the ONNX version to export the model to\n",
+    "                  do_constant_folding=True,  # whether to execute constant folding for optimization\n",
+    "                  input_names=['input'],   # the model's input names\n",
+    "                  output_names=['output'],  # the model's output names\n",
+    "                  dynamic_axes={'input': {0: 'batch_size'},    # variable length axes\n",
+    "                                'output': {0: 'batch_size'}})\n",
+    "\n",
+    "d = ((x).detach().numpy()).reshape([-1]).tolist()\n",
+    "\n",
+    "data = dict(input_shapes=[shape],\n",
+    "            input_data=[d],\n",
+    "            output_data=[((o).detach().numpy()).reshape([-1]).tolist() for o in torch_out])\n",
+    "\n",
+    "# Serialize data into file:\n",
+    "json.dump(data, open(\"input.json\", 'w'))\n",
+    "\n",
+    "\n",
+    "# note that you can also call the following function to generate random data for the model\n",
+    "# it is functionally equivalent to the code above\n",
+    "ezkl.gen_random_data()\n"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "d5e374a2",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "!RUST_LOG=trace\n",
+    "# TODO: Dictionary outputs\n",
+    "res = ezkl.gen_settings(model_path, settings_path)\n",
+    "assert res == True\n"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "cal_path = os.path.join(\"calibration.json\")\n",
+    "\n",
+    "data_array = (torch.randn(20, *shape).detach().numpy()).reshape([-1]).tolist()\n",
+    "\n",
+    "data = dict(input_data = [data_array])\n",
+    "\n",
+    "# Serialize data into file:\n",
+    "json.dump(data, open(cal_path, 'w'))\n",
+    "\n",
+    "res = await ezkl.calibrate_settings(data_path, model_path, settings_path, \"resources\")\n",
+    "assert res == True\n"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "3aa4f090",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "res = ezkl.compile_circuit(model_path, compiled_model_path, settings_path)\n",
+    "assert res == True"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "8b74dcee",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "# srs path\n",
+    "res = await ezkl.get_srs( settings_path)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "18c8b7c7",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "# now generate the witness file \n",
+    "\n",
+    "res = await ezkl.gen_witness(data_path, compiled_model_path, witness_path)\n",
+    "assert os.path.isfile(witness_path)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "b1c561a8",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "\n",
+    "# HERE WE SETUP THE CIRCUIT PARAMS\n",
+    "# WE GOT KEYS\n",
+    "# WE GOT CIRCUIT PARAMETERS\n",
+    "# EVERYTHING ANYONE HAS EVER NEEDED FOR ZK\n",
+    "\n",
+    "\n",
+    "\n",
+    "res = ezkl.setup(\n",
+    "        compiled_model_path,\n",
+    "        vk_path,\n",
+    "        pk_path,\n",
+    "        \n",
+    "    )\n",
+    "\n",
+    "assert res == True\n",
+    "assert os.path.isfile(vk_path)\n",
+    "assert os.path.isfile(pk_path)\n",
+    "assert os.path.isfile(settings_path)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "c384cbc8",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "# GENERATE A PROOF\n",
+    "\n",
+    "\n",
+    "proof_path = os.path.join('test.pf')\n",
+    "\n",
+    "res = ezkl.prove(\n",
+    "        witness_path,\n",
+    "        compiled_model_path,\n",
+    "        pk_path,\n",
+    "        proof_path,\n",
+    "        \n",
+    "        \"single\",\n",
+    "    )\n",
+    "\n",
+    "print(res)\n",
+    "assert os.path.isfile(proof_path)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "76f00d41",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "# VERIFY IT\n",
+    "\n",
+    "res = ezkl.verify(\n",
+    "        proof_path,\n",
+    "        settings_path,\n",
+    "        vk_path,\n",
+    "        \n",
+    "    )\n",
+    "\n",
+    "assert res == True\n",
+    "print(\"verified\")"
+   ]
+  }
+ ],
+ "metadata": {
+  "kernelspec": {
+   "display_name": "Python 3 (ipykernel)",
+   "language": "python",
+   "name": "python3"
+  },
+  "language_info": {
+   "codemirror_mode": {
+    "name": "ipython",
+    "version": 3
+   },
+   "file_extension": ".py",
+   "mimetype": "text/x-python",
+   "name": "python",
+   "nbconvert_exporter": "python",
+   "pygments_lexer": "ipython3",
+   "version": "3.9.15"
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 5
+}

examples/notebooks/logistic_regression.ipynb

@@ -271,7 +271,7 @@
             "name": "python",
             "nbconvert_exporter": "python",
             "pygments_lexer": "ipython3",
-            "version": "3.12.2"
+            "version": "3.12.7"
         }
     },
     "nbformat": 4,

examples/notebooks/mean_postgres.ipynb

@@ -1,456 +1,462 @@
 {
-  "cells": [
-    {
-      "cell_type": "markdown",
-      "metadata": {},
-      "source": [
-        "## Mean of ERC20 transfer amounts\n",
-        "\n",
-        "This notebook shows how to calculate the mean of ERC20 transfer amounts, pulling data in from a Postgres database. First we install and get the necessary libraries running. \n",
-        "The first of which is [shovel](https://indexsupply.com/shovel/docs/#getting-started), which is a library that allows us to pull data from the Ethereum blockchain into a Postgres database.\n",
-        "\n",
-        "Make sure you install postgres if needed https://indexsupply.com/shovel/docs/#getting-started. \n",
-        "\n"
-      ]
-    },
-    {
-      "cell_type": "code",
-      "execution_count": null,
-      "metadata": {},
-      "outputs": [],
-      "source": [
-        "import os\n",
-        "import getpass\n",
-        "import json\n",
-        "import time\n",
-        "import subprocess\n",
-        "\n",
-        "# swap out for the relevant linux/amd64, darwin/arm64, darwin/amd64, windows/amd64\n",
-        "os.system(\"curl -LO https://indexsupply.net/bin/1.0/linux/amd64/shovel\")\n",
-        "os.system(\"chmod +x shovel\")\n",
-        "\n",
-        "\n",
-        "os.environ[\"PG_URL\"] = \"postgres://\" + getpass.getuser() + \":@localhost:5432/shovel\"\n",
-        "\n",
-        "# create a config.json file with the following contents\n",
-        "config = {\n",
-        "  \"pg_url\": \"$PG_URL\",\n",
-        "  \"eth_sources\": [\n",
-        "    {\"name\": \"mainnet\", \"chain_id\": 1, \"url\": \"https://ethereum-rpc.publicnode.com\"},\n",
-        "    {\"name\": \"base\", \"chain_id\": 8453, \"url\": \"https://base-rpc.publicnode.com\"}\n",
-        "  ],\n",
-        "  \"integrations\": [{\n",
-        "    \"name\": \"usdc_transfer\",\n",
-        "    \"enabled\": True,\n",
-        "    \"sources\": [{\"name\": \"mainnet\"}, {\"name\": \"base\"}],\n",
-        "    \"table\": {\n",
-        "      \"name\": \"usdc\",\n",
-        "      \"columns\": [\n",
-        "        {\"name\": \"log_addr\",  \"type\": \"bytea\"},\n",
-        "        {\"name\": \"block_num\", \"type\": \"numeric\"},\n",
-        "        {\"name\": \"f\",         \"type\": \"bytea\"},\n",
-        "        {\"name\": \"t\",         \"type\": \"bytea\"},\n",
-        "        {\"name\": \"v\",         \"type\": \"numeric\"}\n",
-        "      ]\n",
-        "    },\n",
-        "    \"block\": [\n",
-        "      {\"name\": \"block_num\", \"column\": \"block_num\"},\n",
-        "      {\n",
-        "        \"name\": \"log_addr\",\n",
-        "        \"column\": \"log_addr\",\n",
-        "        \"filter_op\": \"contains\",\n",
-        "        \"filter_arg\": [\n",
-        "          \"a0b86991c6218b36c1d19d4a2e9eb0ce3606eb48\",\n",
-        "          \"833589fCD6eDb6E08f4c7C32D4f71b54bdA02913\"\n",
-        "        ]\n",
-        "      }\n",
-        "    ],\n",
-        "    \"event\": {\n",
-        "      \"name\": \"Transfer\",\n",
-        "      \"type\": \"event\",\n",
-        "      \"anonymous\": False,\n",
-        "      \"inputs\": [\n",
-        "        {\"indexed\": True,  \"name\": \"from\",  \"type\": \"address\", \"column\": \"f\"},\n",
-        "        {\"indexed\": True,  \"name\": \"to\",    \"type\": \"address\", \"column\": \"t\"},\n",
-        "        {\"indexed\": False, \"name\": \"value\", \"type\": \"uint256\", \"column\": \"v\"}\n",
-        "      ]\n",
-        "    }\n",
-        "  }]\n",
-        "}\n",
-        "\n",
-        "# write the config to a file\n",
-        "with open(\"config.json\", \"w\") as f:\n",
-        "  f.write(json.dumps(config))\n",
-        "\n",
-        "\n",
-        "# print the two env variables\n",
-        "os.system(\"echo $PG_URL\")\n",
-        "\n",
-        "os.system(\"createdb -h localhost -p 5432 shovel\")\n",
-        "\n",
-        "os.system(\"echo shovel is now installed. starting:\")\n",
-        "\n",
-        "command = [\"./shovel\", \"-config\", \"config.json\"]\n",
-        "proc = subprocess.Popen(command)\n",
-        "\n",
-        "os.system(\"echo shovel started.\")\n",
-        "\n",
-        "time.sleep(10)\n",
-        "\n",
-        "# after we've fetched some data -- kill the process\n",
-        "proc.terminate()\n",
-        "\n"
-      ]
-    },
-    {
-      "cell_type": "markdown",
-      "metadata": {
-        "id": "2wIAHwqH2_mo"
-      },
-      "source": [
-        "**Import Dependencies**"
-      ]
-    },
-    {
-      "cell_type": "code",
-      "execution_count": null,
-      "metadata": {
-        "id": "9Byiv2Nc2MsK"
-      },
-      "outputs": [],
-      "source": [
-        "# check if notebook is in colab\n",
-        "try:\n",
-        "    # install ezkl\n",
-        "    import google.colab\n",
-        "    import subprocess\n",
-        "    import sys\n",
-        "    subprocess.check_call([sys.executable, \"-m\", \"pip\", \"install\", \"ezkl\"])\n",
-        "    subprocess.check_call([sys.executable, \"-m\", \"pip\", \"install\", \"onnx\"])\n",
-        "\n",
-        "# rely on local installation of ezkl if the notebook is not in colab\n",
-        "except:\n",
-        "    pass\n",
-        "\n",
-        "import ezkl\n",
-        "import torch\n",
-        "import datetime\n",
-        "import pandas as pd\n",
-        "import requests\n",
-        "import json\n",
-        "import os\n",
-        "\n",
-        "import logging\n",
-        "# # uncomment for more descriptive logging \n",
-        "FORMAT = '%(levelname)s %(name)s %(asctime)-15s %(filename)s:%(lineno)d %(message)s'\n",
-        "logging.basicConfig(format=FORMAT)\n",
-        "logging.getLogger().setLevel(logging.DEBUG)\n",
-        "\n",
-        "print(\"ezkl version: \", ezkl.__version__)"
-      ]
-    },
-    {
-      "cell_type": "markdown",
-      "metadata": {
-        "id": "osjj-0Ta3E8O"
-      },
-      "source": [
-        "**Create Computational Graph**"
-      ]
-    },
-    {
-      "cell_type": "code",
-      "execution_count": null,
-      "metadata": {
-        "colab": {
-          "base_uri": "https://localhost:8080/"
-        },
-        "id": "x1vl9ZXF3EEW",
-        "outputId": "bda21d02-fe5f-4fb2-8106-f51a8e2e67aa"
-      },
-      "outputs": [],
-      "source": [
-        "from torch import nn\n",
-        "import torch\n",
-        "\n",
-        "\n",
-        "class Model(nn.Module):\n",
-        "    def __init__(self):\n",
-        "        super(Model, self).__init__()\n",
-        "\n",
-        "    # x is a time series \n",
-        "    def forward(self, x):\n",
-        "        return [torch.mean(x)]\n",
-        "\n",
-        "\n",
-        "\n",
-        "\n",
-        "circuit = Model()\n",
-        "\n",
-        "\n",
-        "\n",
-        "\n",
-        "x = 0.1*torch.rand(1,*[1,5], requires_grad=True)\n",
-        "\n",
-        "# # print(torch.__version__)\n",
-        "device = torch.device(\"cuda:0\" if torch.cuda.is_available() else \"cpu\")\n",
-        "\n",
-        "print(device)\n",
-        "\n",
-        "circuit.to(device)\n",
-        "\n",
-        "# Flips the neural net into inference mode\n",
-        "circuit.eval()\n",
-        "\n",
-        "# Export the model\n",
-        "torch.onnx.export(circuit,               # model being run\n",
-        "                      x,                   # model input (or a tuple for multiple inputs)\n",
-        "                      \"lol.onnx\",            # where to save the model (can be a file or file-like object)\n",
-        "                      export_params=True,        # store the trained parameter weights inside the model file\n",
-        "                      opset_version=11,          # the ONNX version to export the model to\n",
-        "                      do_constant_folding=True,  # whether to execute constant folding for optimization\n",
-        "                      input_names = ['input'],   # the model's input names\n",
-        "                      output_names = ['output'], # the model's output names\n",
-        "                      dynamic_axes={'input' : {0 : 'batch_size'},    # variable length axes\n",
-        "                                    'output' : {0 : 'batch_size'}})\n",
-        "\n",
-        "# export(circuit, input_shape=[1, 20])\n",
-        "\n"
-      ]
-    },
-    {
-      "cell_type": "markdown",
-      "metadata": {
-        "id": "E3qCeX-X5xqd"
-      },
-      "source": [
-        "**Set Data Source and Get Data**"
-      ]
-    },
-    {
-      "cell_type": "code",
-      "execution_count": null,
-      "metadata": {
-        "colab": {
-          "base_uri": "https://localhost:8080/"
-        },
-        "id": "6RAMplxk5xPk",
-        "outputId": "bd2158fe-0c00-44fd-e632-6a3f70cdb7c9"
-      },
-      "outputs": [],
-      "source": [
-        "import getpass\n",
-        "# make an input.json file from the df above\n",
-        "input_filename = os.path.join('input.json')\n",
-        "\n",
-        "pg_input_file = dict(input_data = {\n",
-        "  \"host\": \"localhost\",\n",
-        "  # make sure you replace this with your own username\n",
-        "  \"user\": getpass.getuser(),\n",
-        "  \"dbname\": \"shovel\",\n",
-        "  \"password\": \"\",\n",
-        "  \"query\": \"SELECT v FROM usdc ORDER BY block_num DESC LIMIT 5\",\n",
-        "  \"port\": \"5432\",\n",
-        "})\n",
-        "\n",
-        "json_formatted_str = json.dumps(pg_input_file, indent=2)\n",
-        "print(json_formatted_str)\n",
-        "\n",
-        "\n",
-        "    # Serialize data into file:\n",
-        "json.dump(pg_input_file, open(input_filename, 'w' ))\n"
-      ]
-    },
-    {
-      "cell_type": "code",
-      "execution_count": null,
-      "metadata": {},
-      "outputs": [],
-      "source": [
-        "# this corresponds to 4 batches\n",
-        "calibration_filename = os.path.join('calibration.json')\n",
-        "\n",
-        "pg_cal_file = dict(input_data = {\n",
-        "  \"host\": \"localhost\",\n",
-        "  # make sure you replace this with your own username\n",
-        "  \"user\": getpass.getuser(),\n",
-        "  \"dbname\": \"shovel\",\n",
-        "  \"password\": \"\",\n",
-        "  \"query\": \"SELECT v FROM usdc ORDER BY block_num DESC LIMIT 20\",\n",
-        "  \"port\": \"5432\",\n",
-        "})\n",
-        "\n",
-        "    # Serialize data into file:\n",
-        "json.dump( pg_cal_file, open(calibration_filename, 'w' ))"
-      ]
-    },
-    {
-      "cell_type": "markdown",
-      "metadata": {
-        "id": "eLJ7oirQ_HQR"
-      },
-      "source": [
-        "**EZKL Workflow**"
-      ]
-    },
-    {
-      "cell_type": "code",
-      "execution_count": null,
-      "metadata": {
-        "id": "rNw0C9QL6W88"
-      },
-      "outputs": [],
-      "source": [
-        "import subprocess\n",
-        "import os\n",
-        "\n",
-        "onnx_filename = os.path.join('lol.onnx')\n",
-        "compiled_filename = os.path.join('lol.compiled')\n",
-        "settings_filename = os.path.join('settings.json')\n",
-        "\n",
-        "# Generate settings using ezkl\n",
-        "res = ezkl.gen_settings(onnx_filename, settings_filename)\n",
-        "\n",
-        "assert res == True\n",
-        "\n",
-        "res = await ezkl.calibrate_settings(input_filename, onnx_filename, settings_filename, \"resources\")\n",
-        "\n",
-        "assert res == True"
-      ]
-    },
-    {
-      "cell_type": "code",
-      "execution_count": null,
-      "metadata": {},
-      "outputs": [],
-      "source": [
-        "\n",
-        "ezkl.compile_circuit(onnx_filename, compiled_filename, settings_filename)"
-      ]
-    },
-    {
-      "cell_type": "code",
-      "execution_count": null,
-      "metadata": {
-        "colab": {
-          "base_uri": "https://localhost:8080/"
-        },
-        "id": "4MmE9SX66_Il",
-        "outputId": "16403639-66a4-4280-ac7f-6966b75de5a3"
-      },
-      "outputs": [],
-      "source": [
-        "# generate settings\n",
-        "\n",
-        "\n",
-        "# show the settings.json\n",
-        "with open(\"settings.json\") as f:\n",
-        "  data = json.load(f)\n",
-        "  json_formatted_str = json.dumps(data, indent=2)\n",
-        "\n",
-        "  print(json_formatted_str)\n",
-        "\n",
-        "assert os.path.exists(\"settings.json\")\n",
-        "assert os.path.exists(\"input.json\")\n",
-        "assert os.path.exists(\"lol.onnx\")"
-      ]
-    },
-    {
-      "cell_type": "code",
-      "execution_count": null,
-      "metadata": {
-        "id": "fULvvnK7_CMb"
-      },
-      "outputs": [],
-      "source": [
-        "pk_path = os.path.join('test.pk')\n",
-        "vk_path = os.path.join('test.vk')\n",
-        "\n",
-        "\n",
-        "# setup the proof\n",
-        "res = ezkl.setup(\n",
-        "        compiled_filename,\n",
-        "        vk_path,\n",
-        "        pk_path\n",
-        "    )\n",
-        "\n",
-        "assert res == True\n",
-        "assert os.path.isfile(vk_path)\n",
-        "assert os.path.isfile(pk_path)\n",
-        "assert os.path.isfile(settings_filename)\n",
-        "\n"
-      ]
-    },
-    {
-      "cell_type": "code",
-      "execution_count": null,
-      "metadata": {},
-      "outputs": [],
-      "source": [
-        "witness_path = \"witness.json\"\n",
-        "\n",
-        "# generate the witness\n",
-        "res = await ezkl.gen_witness(\n",
-        "        input_filename,\n",
-        "        compiled_filename,\n",
-        "        witness_path\n",
-        "    )\n"
-      ]
-    },
-    {
-      "cell_type": "code",
-      "execution_count": null,
-      "metadata": {
-        "colab": {
-          "base_uri": "https://localhost:8080/"
-        },
-        "id": "Oog3j6Kd-Wed",
-        "outputId": "5839d0c1-5b43-476e-c2f8-6707de562260"
-      },
-      "outputs": [],
-      "source": [
-        "# prove the zk circuit\n",
-        "# GENERATE A PROOF\n",
-        "proof_path = os.path.join('test.pf')\n",
-        "\n",
-        "\n",
-        "proof = ezkl.prove(\n",
-        "        witness_path,\n",
-        "        compiled_filename,\n",
-        "        pk_path,\n",
-        "        proof_path,\n",
-        "        \"single\"\n",
-        "    )\n",
-        "\n",
-        "\n",
-        "print(\"proved\")\n",
-        "\n",
-        "assert os.path.isfile(proof_path)\n",
-        "\n"
-      ]
-    }
-  ],
-  "metadata": {
-    "colab": {
-      "provenance": []
-    },
-    "kernelspec": {
-      "display_name": "Python 3",
-      "name": "python3"
-    },
-    "language_info": {
-      "codemirror_mode": {
-        "name": "ipython",
-        "version": 3
-      },
-      "file_extension": ".py",
-      "mimetype": "text/x-python",
-      "name": "python",
-      "nbconvert_exporter": "python",
-      "pygments_lexer": "ipython3",
-      "version": "3.12.2"
-    }
+ "cells": [
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "## Mean of ERC20 transfer amounts\n",
+    "\n",
+    "This notebook shows how to calculate the mean of ERC20 transfer amounts, pulling data in from a Postgres database. First we install and get the necessary libraries running. \n",
+    "The first of which is [shovel](https://indexsupply.com/shovel/docs/#getting-started), which is a library that allows us to pull data from the Ethereum blockchain into a Postgres database.\n",
+    "\n",
+    "Make sure you install postgres if needed https://indexsupply.com/shovel/docs/#getting-started. \n",
+    "\n"
+   ]
   },
-  "nbformat": 4,
-  "nbformat_minor": 0
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "import os\n",
+    "import getpass\n",
+    "import json\n",
+    "import time\n",
+    "import subprocess\n",
+    "\n",
+    "# swap out for the relevant linux/amd64, darwin/arm64, darwin/amd64, windows/amd64\n",
+    "os.system(\"curl -LO https://indexsupply.net/bin/1.0/linux/amd64/shovel\")\n",
+    "os.system(\"chmod +x shovel\")\n",
+    "\n",
+    "\n",
+    "os.environ[\"PG_URL\"] = \"postgres://\" + getpass.getuser() + \":@localhost:5432/shovel\"\n",
+    "\n",
+    "# create a config.json file with the following contents\n",
+    "config = {\n",
+    "  \"pg_url\": \"$PG_URL\",\n",
+    "  \"eth_sources\": [\n",
+    "    {\"name\": \"mainnet\", \"chain_id\": 1, \"url\": \"https://ethereum-rpc.publicnode.com\"},\n",
+    "    {\"name\": \"base\", \"chain_id\": 8453, \"url\": \"https://base-rpc.publicnode.com\"}\n",
+    "  ],\n",
+    "  \"integrations\": [{\n",
+    "    \"name\": \"usdc_transfer\",\n",
+    "    \"enabled\": True,\n",
+    "    \"sources\": [{\"name\": \"mainnet\"}, {\"name\": \"base\"}],\n",
+    "    \"table\": {\n",
+    "      \"name\": \"usdc\",\n",
+    "      \"columns\": [\n",
+    "        {\"name\": \"log_addr\",  \"type\": \"bytea\"},\n",
+    "        {\"name\": \"block_num\", \"type\": \"numeric\"},\n",
+    "        {\"name\": \"f\",         \"type\": \"bytea\"},\n",
+    "        {\"name\": \"t\",         \"type\": \"bytea\"},\n",
+    "        {\"name\": \"v\",         \"type\": \"numeric\"}\n",
+    "      ]\n",
+    "    },\n",
+    "    \"block\": [\n",
+    "      {\"name\": \"block_num\", \"column\": \"block_num\"},\n",
+    "      {\n",
+    "        \"name\": \"log_addr\",\n",
+    "        \"column\": \"log_addr\",\n",
+    "        \"filter_op\": \"contains\",\n",
+    "        \"filter_arg\": [\n",
+    "          \"a0b86991c6218b36c1d19d4a2e9eb0ce3606eb48\",\n",
+    "          \"833589fCD6eDb6E08f4c7C32D4f71b54bdA02913\"\n",
+    "        ]\n",
+    "      }\n",
+    "    ],\n",
+    "    \"event\": {\n",
+    "      \"name\": \"Transfer\",\n",
+    "      \"type\": \"event\",\n",
+    "      \"anonymous\": False,\n",
+    "      \"inputs\": [\n",
+    "        {\"indexed\": True,  \"name\": \"from\",  \"type\": \"address\", \"column\": \"f\"},\n",
+    "        {\"indexed\": True,  \"name\": \"to\",    \"type\": \"address\", \"column\": \"t\"},\n",
+    "        {\"indexed\": False, \"name\": \"value\", \"type\": \"uint256\", \"column\": \"v\"}\n",
+    "      ]\n",
+    "    }\n",
+    "  }]\n",
+    "}\n",
+    "\n",
+    "# write the config to a file\n",
+    "with open(\"config.json\", \"w\") as f:\n",
+    "  f.write(json.dumps(config))\n",
+    "\n",
+    "\n",
+    "# print the two env variables\n",
+    "os.system(\"echo $PG_URL\")\n",
+    "\n",
+    "os.system(\"createdb -h localhost -p 5432 shovel\")\n",
+    "\n",
+    "os.system(\"echo shovel is now installed. starting:\")\n",
+    "\n",
+    "command = [\"./shovel\", \"-config\", \"config.json\"]\n",
+    "proc = subprocess.Popen(command)\n",
+    "\n",
+    "os.system(\"echo shovel started.\")\n",
+    "\n",
+    "time.sleep(10)\n",
+    "\n",
+    "# after we've fetched some data -- kill the process\n",
+    "proc.terminate()\n",
+    "\n"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {
+    "id": "2wIAHwqH2_mo"
+   },
+   "source": [
+    "**Import Dependencies**"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {
+    "id": "9Byiv2Nc2MsK"
+   },
+   "outputs": [],
+   "source": [
+    "# check if notebook is in colab\n",
+    "try:\n",
+    "    # install ezkl\n",
+    "    import google.colab\n",
+    "    import subprocess\n",
+    "    import sys\n",
+    "    subprocess.check_call([sys.executable, \"-m\", \"pip\", \"install\", \"ezkl\"])\n",
+    "    subprocess.check_call([sys.executable, \"-m\", \"pip\", \"install\", \"onnx\"])\n",
+    "\n",
+    "# rely on local installation of ezkl if the notebook is not in colab\n",
+    "except:\n",
+    "    pass\n",
+    "\n",
+    "import ezkl\n",
+    "import torch\n",
+    "import datetime\n",
+    "import pandas as pd\n",
+    "import requests\n",
+    "import json\n",
+    "import os\n",
+    "\n",
+    "import logging\n",
+    "# # uncomment for more descriptive logging \n",
+    "FORMAT = '%(levelname)s %(name)s %(asctime)-15s %(filename)s:%(lineno)d %(message)s'\n",
+    "logging.basicConfig(format=FORMAT)\n",
+    "logging.getLogger().setLevel(logging.DEBUG)\n",
+    "\n",
+    "print(\"ezkl version: \", ezkl.__version__)"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {
+    "id": "osjj-0Ta3E8O"
+   },
+   "source": [
+    "**Create Computational Graph**"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {
+    "colab": {
+     "base_uri": "https://localhost:8080/"
+    },
+    "id": "x1vl9ZXF3EEW",
+    "outputId": "bda21d02-fe5f-4fb2-8106-f51a8e2e67aa"
+   },
+   "outputs": [],
+   "source": [
+    "from torch import nn\n",
+    "import torch\n",
+    "\n",
+    "\n",
+    "class Model(nn.Module):\n",
+    "    def __init__(self):\n",
+    "        super(Model, self).__init__()\n",
+    "\n",
+    "    # x is a time series \n",
+    "    def forward(self, x):\n",
+    "        return [torch.mean(x)]\n",
+    "\n",
+    "\n",
+    "\n",
+    "\n",
+    "circuit = Model()\n",
+    "\n",
+    "\n",
+    "\n",
+    "\n",
+    "x = 0.1*torch.rand(1,*[1,5], requires_grad=True)\n",
+    "\n",
+    "# # print(torch.__version__)\n",
+    "device = torch.device(\"cuda:0\" if torch.cuda.is_available() else \"cpu\")\n",
+    "\n",
+    "print(device)\n",
+    "\n",
+    "circuit.to(device)\n",
+    "\n",
+    "# Flips the neural net into inference mode\n",
+    "circuit.eval()\n",
+    "\n",
+    "# Export the model\n",
+    "torch.onnx.export(circuit,               # model being run\n",
+    "                      x,                   # model input (or a tuple for multiple inputs)\n",
+    "                      \"lol.onnx\",            # where to save the model (can be a file or file-like object)\n",
+    "                      export_params=True,        # store the trained parameter weights inside the model file\n",
+    "                      opset_version=11,          # the ONNX version to export the model to\n",
+    "                      do_constant_folding=True,  # whether to execute constant folding for optimization\n",
+    "                      input_names = ['input'],   # the model's input names\n",
+    "                      output_names = ['output'], # the model's output names\n",
+    "                      dynamic_axes={'input' : {0 : 'batch_size'},    # variable length axes\n",
+    "                                    'output' : {0 : 'batch_size'}})\n",
+    "\n",
+    "# export(circuit, input_shape=[1, 20])\n",
+    "\n"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {
+    "id": "E3qCeX-X5xqd"
+   },
+   "source": [
+    "**Set Data Source and Get Data**"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {
+    "colab": {
+     "base_uri": "https://localhost:8080/"
+    },
+    "id": "6RAMplxk5xPk",
+    "outputId": "bd2158fe-0c00-44fd-e632-6a3f70cdb7c9"
+   },
+   "outputs": [],
+   "source": [
+    "import getpass\n",
+    "# make an input.json file from the df above\n",
+    "input_filename = os.path.join('input.json')\n",
+    "\n",
+    "pg_input_file = dict(input_data = {\n",
+    "  \"host\": \"localhost\",\n",
+    "  # make sure you replace this with your own username\n",
+    "  \"user\": getpass.getuser(),\n",
+    "  \"dbname\": \"shovel\",\n",
+    "  \"password\": \"\",\n",
+    "  \"query\": \"SELECT v FROM usdc ORDER BY block_num DESC LIMIT 5\",\n",
+    "  \"port\": \"5432\",\n",
+    "})\n",
+    "\n",
+    "json_formatted_str = json.dumps(pg_input_file, indent=2)\n",
+    "print(json_formatted_str)\n",
+    "\n",
+    "\n",
+    "    # Serialize data into file:\n",
+    "json.dump(pg_input_file, open(input_filename, 'w' ))\n"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "# this corresponds to 4 batches\n",
+    "calibration_filename = os.path.join('calibration.json')\n",
+    "\n",
+    "pg_cal_file = dict(input_data = {\n",
+    "  \"host\": \"localhost\",\n",
+    "  # make sure you replace this with your own username\n",
+    "  \"user\": getpass.getuser(),\n",
+    "  \"dbname\": \"shovel\",\n",
+    "  \"password\": \"\",\n",
+    "  \"query\": \"SELECT v FROM usdc ORDER BY block_num DESC LIMIT 20\",\n",
+    "  \"port\": \"5432\",\n",
+    "})\n",
+    "\n",
+    "    # Serialize data into file:\n",
+    "json.dump( pg_cal_file, open(calibration_filename, 'w' ))"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {
+    "id": "eLJ7oirQ_HQR"
+   },
+   "source": [
+    "**EZKL Workflow**"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {
+    "id": "rNw0C9QL6W88"
+   },
+   "outputs": [],
+   "source": [
+    "import subprocess\n",
+    "import os\n",
+    "\n",
+    "onnx_filename = os.path.join('lol.onnx')\n",
+    "compiled_filename = os.path.join('lol.compiled')\n",
+    "settings_filename = os.path.join('settings.json')\n",
+    "\n",
+    "run_args = ezkl.PyRunArgs()\n",
+    "run_args.decomp_legs = 4\n",
+    "\n",
+    "# Generate settings using ezkl\n",
+    "res = ezkl.gen_settings(onnx_filename, settings_filename, py_run_args=run_args)\n",
+    "\n",
+    "assert res == True\n",
+    "\n",
+    "res = await ezkl.calibrate_settings(input_filename, onnx_filename, settings_filename, \"resources\")\n",
+    "\n",
+    "assert res == True\n",
+    "\n",
+    "await ezkl.get_srs(settings_filename)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "\n",
+    "ezkl.compile_circuit(onnx_filename, compiled_filename, settings_filename)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {
+    "colab": {
+     "base_uri": "https://localhost:8080/"
+    },
+    "id": "4MmE9SX66_Il",
+    "outputId": "16403639-66a4-4280-ac7f-6966b75de5a3"
+   },
+   "outputs": [],
+   "source": [
+    "# generate settings\n",
+    "\n",
+    "\n",
+    "# show the settings.json\n",
+    "with open(\"settings.json\") as f:\n",
+    "  data = json.load(f)\n",
+    "  json_formatted_str = json.dumps(data, indent=2)\n",
+    "\n",
+    "  print(json_formatted_str)\n",
+    "\n",
+    "assert os.path.exists(\"settings.json\")\n",
+    "assert os.path.exists(\"input.json\")\n",
+    "assert os.path.exists(\"lol.onnx\")"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {
+    "id": "fULvvnK7_CMb"
+   },
+   "outputs": [],
+   "source": [
+    "pk_path = os.path.join('test.pk')\n",
+    "vk_path = os.path.join('test.vk')\n",
+    "\n",
+    "\n",
+    "# setup the proof\n",
+    "res = ezkl.setup(\n",
+    "        compiled_filename,\n",
+    "        vk_path,\n",
+    "        pk_path\n",
+    "    )\n",
+    "\n",
+    "assert res == True\n",
+    "assert os.path.isfile(vk_path)\n",
+    "assert os.path.isfile(pk_path)\n",
+    "assert os.path.isfile(settings_filename)\n",
+    "\n"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "witness_path = \"witness.json\"\n",
+    "\n",
+    "# generate the witness\n",
+    "res = await ezkl.gen_witness(\n",
+    "        input_filename,\n",
+    "        compiled_filename,\n",
+    "        witness_path\n",
+    "    )\n"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {
+    "colab": {
+     "base_uri": "https://localhost:8080/"
+    },
+    "id": "Oog3j6Kd-Wed",
+    "outputId": "5839d0c1-5b43-476e-c2f8-6707de562260"
+   },
+   "outputs": [],
+   "source": [
+    "# prove the zk circuit\n",
+    "# GENERATE A PROOF\n",
+    "proof_path = os.path.join('test.pf')\n",
+    "\n",
+    "\n",
+    "proof = ezkl.prove(\n",
+    "        witness_path,\n",
+    "        compiled_filename,\n",
+    "        pk_path,\n",
+    "        proof_path,\n",
+    "        \"single\"\n",
+    "    )\n",
+    "\n",
+    "\n",
+    "print(\"proved\")\n",
+    "\n",
+    "assert os.path.isfile(proof_path)\n",
+    "\n"
+   ]
+  }
+ ],
+ "metadata": {
+  "colab": {
+   "provenance": []
+  },
+  "kernelspec": {
+   "display_name": ".env",
+   "language": "python",
+   "name": "python3"
+  },
+  "language_info": {
+   "codemirror_mode": {
+    "name": "ipython",
+    "version": 3
+   },
+   "file_extension": ".py",
+   "mimetype": "text/x-python",
+   "name": "python",
+   "nbconvert_exporter": "python",
+   "pygments_lexer": "ipython3",
+   "version": "3.12.7"
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 0
 }

examples/notebooks/neural_bow.ipynb

@@ -0,0 +1,766 @@
+{
+ "cells": [
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "\n",
+    "This is a zk version of the tutorial found [here](https://github.com/bentrevett/pytorch-sentiment-analysis/blob/main/1%20-%20Neural%20Bag%20of%20Words.ipynb). The original tutorial is part of the PyTorch Sentiment Analysis series by Ben Trevett.\n",
+    "\n",
+    "1 - NBoW\n",
+    "\n",
+    "In this series we'll be building a machine learning model to perform sentiment analysis -- a subset of text classification where the task is to detect if a given sentence is positive or negative -- using PyTorch and torchtext. The dataset used will be movie reviews from the IMDb dataset, which we'll obtain using the datasets library.\n"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "\n",
+    "Preparing Data\n",
+    "\n",
+    "Before we can implement our NBoW model, we first have to perform quite a few steps to get our data ready to use. NLP usually requires quite a lot of data wrangling beforehand, though libraries such as datasets and torchtext handle most of this for us.\n",
+    "\n",
+    "The steps to take are:\n",
+    "\n",
+    "    1. importing modules\n",
+    "    2. loading data\n",
+    "    3. tokenizing data\n",
+    "    4. creating data splits\n",
+    "    5. creating a vocabulary\n",
+    "    6. numericalizing data\n",
+    "    7. creating the data loaders\n"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "! pip install torchtex"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "import collections\n",
+    "\n",
+    "import datasets\n",
+    "import matplotlib.pyplot as plt\n",
+    "import numpy as np\n",
+    "import torch\n",
+    "import torch.nn as nn\n",
+    "import torch.optim as optim\n",
+    "import torchtext\n",
+    "import tqdm\n",
+    "\n",
+    "# It is usually good practice to run your experiments multiple times with different random seeds -- both to measure the variance of your model and also to avoid having results only calculated with either \"good\" or \"bad\" seeds, i.e. being very lucky or unlucky with the randomness in the training process.\n",
+    "\n",
+    "seed = 1234\n",
+    "\n",
+    "np.random.seed(seed)\n",
+    "torch.manual_seed(seed)\n",
+    "torch.cuda.manual_seed(seed)\n",
+    "torch.backends.cudnn.deterministic = True\n"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "train_data, test_data = datasets.load_dataset(\"imdb\", split=[\"train\", \"test\"])"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "We can check the features attribute of a split to get more information about the features. We can see that text is a Value of dtype=string -- in other words, it's a string -- and that label is a ClassLabel. A ClassLabel means the feature is an integer representation of which class the example belongs to. num_classes=2 means that our labels are one of two values, 0 or 1, and names=['neg', 'pos'] gives us the human-readable versions of those values. Thus, a label of 0 means the example is a negative review and a label of 1 means the example is a positive review."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "train_data.features\n"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "train_data[0]"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "One of the first things we need to do to our data is tokenize it. Machine learning models aren't designed to handle strings, they're design to handle numbers. So what we need to do is break down our string into individual tokens, and then convert these tokens to numbers. We'll get to the conversion later, but first we'll look at tokenization.\n",
+    "\n",
+    "Tokenization involves using a tokenizer to process the strings in our dataset. A tokenizer is a function that goes from a string to a list of strings. There are many types of tokenizers available, but we're going to use a relatively simple one provided by torchtext called the basic_english tokenizer. We load our tokenizer as such:"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "tokenizer = torchtext.data.utils.get_tokenizer(\"basic_english\")"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "def tokenize_example(example, tokenizer, max_length):\n",
+    "    tokens = tokenizer(example[\"text\"])[:max_length]\n",
+    "    return {\"tokens\": tokens}\n",
+    "\n",
+    "\n",
+    "max_length = 256\n",
+    "\n",
+    "train_data = train_data.map(\n",
+    "    tokenize_example, fn_kwargs={\"tokenizer\": tokenizer, \"max_length\": max_length}\n",
+    ")\n",
+    "test_data = test_data.map(\n",
+    "    tokenize_example, fn_kwargs={\"tokenizer\": tokenizer, \"max_length\": max_length}\n",
+    ")\n",
+    "\n",
+    "\n",
+    "# create validation data \n",
+    "# Why have both a validation set and a test set? Your test set respresents the real world data that you'd see if you actually deployed this model. You won't be able to see what data your model will be fed once deployed, and your test set is supposed to reflect that. Every time we tune our model hyperparameters or training set-up to make it do a bit better on the test set, we are leak information from the test set into the training process. If we do this too often then we begin to overfit on the test set. Hence, we need some data which can act as a \"proxy\" test set which we can look at more frequently in order to evaluate how well our model actually does on unseen data -- this is the validation set.\n",
+    "\n",
+    "test_size = 0.25\n",
+    "\n",
+    "train_valid_data = train_data.train_test_split(test_size=test_size)\n",
+    "train_data = train_valid_data[\"train\"]\n",
+    "valid_data = train_valid_data[\"test\"]"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "Next, we have to build a vocabulary. This is look-up table where every unique token in your dataset has a corresponding index (an integer).\n",
+    "\n",
+    "We do this as machine learning models cannot operate on strings, only numerical vaslues. Each index is used to construct a one-hot vector for each token. A one-hot vector is a vector where all the elements are 0, except one, which is 1, and the dimensionality is the total number of unique tokens in your vocabulary, commonly denoted by V."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "min_freq = 5\n",
+    "special_tokens = [\"<unk>\", \"<pad>\"]\n",
+    "\n",
+    "vocab = torchtext.vocab.build_vocab_from_iterator(\n",
+    "    train_data[\"tokens\"],\n",
+    "    min_freq=min_freq,\n",
+    "    specials=special_tokens,\n",
+    ")\n",
+    "\n",
+    "# We store the indices of the unknown and padding tokens (zero and one, respectively) in variables, as we'll use these further on in this notebook.\n",
+    "\n",
+    "unk_index = vocab[\"<unk>\"]\n",
+    "pad_index = vocab[\"<pad>\"]\n",
+    "\n",
+    "\n",
+    "vocab.set_default_index(unk_index)\n",
+    "\n",
+    "# To look-up a list of tokens, we can use the vocabulary's lookup_indices method.\n",
+    "vocab.lookup_indices([\"hello\", \"world\", \"some_token\", \"<pad>\"])"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "Now we have our vocabulary, we can numericalize our data. This involves converting the tokens within our dataset into indices. Similar to how we tokenized our data using the Dataset.map method, we'll define a function that takes an example and our vocabulary, gets the index for each token in each example and then creates an ids field which containes the numericalized tokens."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "def numericalize_example(example, vocab):\n",
+    "    ids = vocab.lookup_indices(example[\"tokens\"])\n",
+    "    return {\"ids\": ids}\n",
+    "\n",
+    "train_data = train_data.map(numericalize_example, fn_kwargs={\"vocab\": vocab})\n",
+    "valid_data = valid_data.map(numericalize_example, fn_kwargs={\"vocab\": vocab})\n",
+    "test_data = test_data.map(numericalize_example, fn_kwargs={\"vocab\": vocab})\n",
+    "\n",
+    "train_data = train_data.with_format(type=\"torch\", columns=[\"ids\", \"label\"])\n",
+    "valid_data = valid_data.with_format(type=\"torch\", columns=[\"ids\", \"label\"])\n",
+    "test_data = test_data.with_format(type=\"torch\", columns=[\"ids\", \"label\"])\n"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "The final step of preparing the data is creating the data loaders. We can iterate over a data loader to retrieve batches of examples. This is also where we will perform any padding that is necessary.\n",
+    "\n",
+    "We first need to define a function to collate a batch, consisting of a list of examples, into what we want our data loader to output.\n",
+    "\n",
+    "Here, our desired output from the data loader is a dictionary with keys of \"ids\" and \"label\".\n",
+    "\n",
+    "The value of batch[\"ids\"] should be a tensor of shape [batch size, length], where length is the length of the longest sentence (in terms of tokens) within the batch, and all sentences shorter than this should be padded to that length.\n",
+    "\n",
+    "The value of batch[\"label\"] should be a tensor of shape [batch size] consisting of the label for each sentence in the batch.\n",
+    "\n",
+    "We define a function, get_collate_fn, which is passed the pad token index and returns the actual collate function. Within the actual collate function, collate_fn, we get a list of \"ids\" tensors for each example in the batch, and then use the pad_sequence function, which converts the list of tensors into the desired [batch size, length] shaped tensor and performs padding using the specified pad_index. By default, pad_sequence will return a [length, batch size] shaped tensor, but by setting batch_first=True, these two dimensions are switched. We get a list of \"label\" tensors and convert the list of tensors into a single [batch size] shaped tensor."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "def get_collate_fn(pad_index):\n",
+    "    def collate_fn(batch):\n",
+    "        batch_ids = [i[\"ids\"] for i in batch]\n",
+    "        batch_ids = nn.utils.rnn.pad_sequence(\n",
+    "            batch_ids, padding_value=pad_index, batch_first=True\n",
+    "        )\n",
+    "        batch_label = [i[\"label\"] for i in batch]\n",
+    "        batch_label = torch.stack(batch_label)\n",
+    "        batch = {\"ids\": batch_ids, \"label\": batch_label}\n",
+    "        return batch\n",
+    "\n",
+    "    return collate_fn\n",
+    "\n",
+    "def get_data_loader(dataset, batch_size, pad_index, shuffle=False):\n",
+    "    collate_fn = get_collate_fn(pad_index)\n",
+    "    data_loader = torch.utils.data.DataLoader(\n",
+    "        dataset=dataset,\n",
+    "        batch_size=batch_size,\n",
+    "        collate_fn=collate_fn,\n",
+    "        shuffle=shuffle,\n",
+    "    )\n",
+    "    return data_loader\n",
+    "\n",
+    "\n",
+    "batch_size = 512\n",
+    "\n",
+    "train_data_loader = get_data_loader(train_data, batch_size, pad_index, shuffle=True)\n",
+    "valid_data_loader = get_data_loader(valid_data, batch_size, pad_index)\n",
+    "test_data_loader = get_data_loader(test_data, batch_size, pad_index)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "\n",
+    "\n",
+    "class NBoW(nn.Module):\n",
+    "    def __init__(self, vocab_size, embedding_dim, output_dim, pad_index):\n",
+    "        super().__init__()\n",
+    "        self.embedding = nn.Embedding(vocab_size, embedding_dim, padding_idx=pad_index)\n",
+    "        self.fc = nn.Linear(embedding_dim, output_dim)\n",
+    "\n",
+    "    def forward(self, ids):\n",
+    "        # ids = [batch size, seq len]\n",
+    "        embedded = self.embedding(ids)\n",
+    "        # embedded = [batch size, seq len, embedding dim]\n",
+    "        pooled = embedded.mean(dim=1)\n",
+    "        # pooled = [batch size, embedding dim]\n",
+    "        prediction = self.fc(pooled)\n",
+    "        # prediction = [batch size, output dim]\n",
+    "        return prediction\n",
+    "\n",
+    "\n",
+    "vocab_size = len(vocab)\n",
+    "embedding_dim = 300\n",
+    "output_dim = len(train_data.unique(\"label\"))\n",
+    "\n",
+    "model = NBoW(vocab_size, embedding_dim, output_dim, pad_index)\n",
+    "\n",
+    "def count_parameters(model):\n",
+    "    return sum(p.numel() for p in model.parameters() if p.requires_grad)\n",
+    "\n",
+    "\n",
+    "print(f\"The model has {count_parameters(model):,} trainable parameters\")\n",
+    "\n",
+    "vectors = torchtext.vocab.GloVe()\n",
+    "\n",
+    "pretrained_embedding = vectors.get_vecs_by_tokens(vocab.get_itos())\n",
+    "\n",
+    "optimizer = optim.Adam(model.parameters())\n",
+    "\n",
+    "criterion = nn.CrossEntropyLoss()\n",
+    "\n",
+    "device = torch.device(\"cuda\" if torch.cuda.is_available() else \"cpu\")\n",
+    "\n",
+    "model = model.to(device)\n",
+    "criterion = criterion.to(device)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "def train(data_loader, model, criterion, optimizer, device):\n",
+    "    model.train()\n",
+    "    epoch_losses = []\n",
+    "    epoch_accs = []\n",
+    "    for batch in tqdm.tqdm(data_loader, desc=\"training...\"):\n",
+    "        ids = batch[\"ids\"].to(device)\n",
+    "        label = batch[\"label\"].to(device)\n",
+    "        prediction = model(ids)\n",
+    "        loss = criterion(prediction, label)\n",
+    "        accuracy = get_accuracy(prediction, label)\n",
+    "        optimizer.zero_grad()\n",
+    "        loss.backward()\n",
+    "        optimizer.step()\n",
+    "        epoch_losses.append(loss.item())\n",
+    "        epoch_accs.append(accuracy.item())\n",
+    "    return np.mean(epoch_losses), np.mean(epoch_accs)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "def evaluate(data_loader, model, criterion, device):\n",
+    "    model.eval()\n",
+    "    epoch_losses = []\n",
+    "    epoch_accs = []\n",
+    "    with torch.no_grad():\n",
+    "        for batch in tqdm.tqdm(data_loader, desc=\"evaluating...\"):\n",
+    "            ids = batch[\"ids\"].to(device)\n",
+    "            label = batch[\"label\"].to(device)\n",
+    "            prediction = model(ids)\n",
+    "            loss = criterion(prediction, label)\n",
+    "            accuracy = get_accuracy(prediction, label)\n",
+    "            epoch_losses.append(loss.item())\n",
+    "            epoch_accs.append(accuracy.item())\n",
+    "    return np.mean(epoch_losses), np.mean(epoch_accs)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "def get_accuracy(prediction, label):\n",
+    "    batch_size, _ = prediction.shape\n",
+    "    predicted_classes = prediction.argmax(dim=-1)\n",
+    "    correct_predictions = predicted_classes.eq(label).sum()\n",
+    "    accuracy = correct_predictions / batch_size\n",
+    "    return accuracy"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "n_epochs = 10\n",
+    "best_valid_loss = float(\"inf\")\n",
+    "\n",
+    "metrics = collections.defaultdict(list)\n",
+    "\n",
+    "for epoch in range(n_epochs):\n",
+    "    train_loss, train_acc = train(\n",
+    "        train_data_loader, model, criterion, optimizer, device\n",
+    "    )\n",
+    "    valid_loss, valid_acc = evaluate(valid_data_loader, model, criterion, device)\n",
+    "    metrics[\"train_losses\"].append(train_loss)\n",
+    "    metrics[\"train_accs\"].append(train_acc)\n",
+    "    metrics[\"valid_losses\"].append(valid_loss)\n",
+    "    metrics[\"valid_accs\"].append(valid_acc)\n",
+    "    if valid_loss < best_valid_loss:\n",
+    "        best_valid_loss = valid_loss\n",
+    "        torch.save(model.state_dict(), \"nbow.pt\")\n",
+    "    print(f\"epoch: {epoch}\")\n",
+    "    print(f\"train_loss: {train_loss:.3f}, train_acc: {train_acc:.3f}\")\n",
+    "    print(f\"valid_loss: {valid_loss:.3f}, valid_acc: {valid_acc:.3f}\")"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "fig = plt.figure(figsize=(10, 6))\n",
+    "ax = fig.add_subplot(1, 1, 1)\n",
+    "ax.plot(metrics[\"train_losses\"], label=\"train loss\")\n",
+    "ax.plot(metrics[\"valid_losses\"], label=\"valid loss\")\n",
+    "ax.set_xlabel(\"epoch\")\n",
+    "ax.set_ylabel(\"loss\")\n",
+    "ax.set_xticks(range(n_epochs))\n",
+    "ax.legend()\n",
+    "ax.grid()"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "fig = plt.figure(figsize=(10, 6))\n",
+    "ax = fig.add_subplot(1, 1, 1)\n",
+    "ax.plot(metrics[\"train_accs\"], label=\"train accuracy\")\n",
+    "ax.plot(metrics[\"valid_accs\"], label=\"valid accuracy\")\n",
+    "ax.set_xlabel(\"epoch\")\n",
+    "ax.set_ylabel(\"loss\")\n",
+    "ax.set_xticks(range(n_epochs))\n",
+    "ax.legend()\n",
+    "ax.grid()"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "model.load_state_dict(torch.load(\"nbow.pt\"))\n",
+    "\n",
+    "test_loss, test_acc = evaluate(test_data_loader, model, criterion, device)\n",
+    "\n",
+    "print(f\"test_loss: {test_loss:.3f}, test_acc: {test_acc:.3f}\")"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "def predict_sentiment(text, model, tokenizer, vocab, device):\n",
+    "    tokens = tokenizer(text)\n",
+    "    ids = vocab.lookup_indices(tokens)\n",
+    "    tensor = torch.LongTensor(ids).unsqueeze(dim=0).to(device)\n",
+    "    prediction = model(tensor).squeeze(dim=0)\n",
+    "    probability = torch.softmax(prediction, dim=-1)\n",
+    "    predicted_class = prediction.argmax(dim=-1).item()\n",
+    "    predicted_probability = probability[predicted_class].item()\n",
+    "    return predicted_class, predicted_probability"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "text = \"This film is terrible!\"\n",
+    "\n",
+    "predict_sentiment(text, model, tokenizer, vocab, device)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "text = \"This film is great!\"\n",
+    "\n",
+    "predict_sentiment(text, model, tokenizer, vocab, device)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "text = \"This film is not terrible, it's great!\"\n",
+    "\n",
+    "predict_sentiment(text, model, tokenizer, vocab, device)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "text = \"This film is not great, it's terrible!\"\n",
+    "\n",
+    "predict_sentiment(text, model, tokenizer, vocab, device)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "def text_to_tensor(text, tokenizer, vocab, device):\n",
+    "    tokens = tokenizer(text)\n",
+    "    ids = vocab.lookup_indices(tokens)\n",
+    "    tensor = torch.LongTensor(ids).unsqueeze(dim=0).to(device)\n",
+    "    return tensor\n",
+    "\n"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "Now we do onnx stuff to get the data ready for the zk-circuit."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "\n",
+    "import json\n",
+    "\n",
+    "text = \"This film is terrible!\"\n",
+    "x = text_to_tensor(text, tokenizer, vocab, device)\n",
+    "\n",
+    "# Flips the neural net into inference mode\n",
+    "model.eval()\n",
+    "model.to('cpu')\n",
+    "\n",
+    "model_path = \"network.onnx\"\n",
+    "data_path = \"input.json\"\n",
+    "\n",
+    "    # Export the model\n",
+    "torch.onnx.export(model,               # model being run\n",
+    "                      x,                   # model input (or a tuple for multiple inputs)\n",
+    "                      model_path,            # where to save the model (can be a file or file-like object)\n",
+    "                      export_params=True,        # store the trained parameter weights inside the model file\n",
+    "                      opset_version=10,          # the ONNX version to export the model to\n",
+    "                      do_constant_folding=True,  # whether to execute constant folding for optimization\n",
+    "                      input_names = ['input'],   # the model's input names\n",
+    "                      output_names = ['output'], # the model's output names\n",
+    "                      dynamic_axes={'input' : {0 : 'batch_size'},    # variable length axes\n",
+    "                                    'output' : {0 : 'batch_size'}})\n",
+    "\n",
+    "\n",
+    "\n",
+    "data_array = ((x).detach().numpy()).reshape([-1]).tolist()\n",
+    "\n",
+    "data_json = dict(input_data = [data_array])\n",
+    "\n",
+    "print(data_json)\n",
+    "\n",
+    "    # Serialize data into file:\n",
+    "json.dump(data_json, open(data_path, 'w'))\n"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "import ezkl\n",
+    "\n",
+    "run_args = ezkl.PyRunArgs()\n",
+    "run_args.logrows = 23\n",
+    "run_args.scale_rebase_multiplier = 10\n",
+    "# inputs should be auditable by all\n",
+    "run_args.input_visibility = \"public\"\n",
+    "# same with outputs\n",
+    "run_args.output_visibility = \"public\"\n",
+    "# for simplicity, we'll just use the fixed model visibility: i.e it is public and can't be changed by the prover\n",
+    "run_args.param_visibility = \"fixed\"\n",
+    "\n",
+    "\n",
+    "# TODO: Dictionary outputs\n",
+    "res = ezkl.gen_settings(py_run_args=run_args)\n",
+    "assert res == True\n",
+    "\n"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "res = ezkl.compile_circuit()\n",
+    "assert res == True"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "# srs path\n",
+    "res = await ezkl.get_srs()"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "# now generate the witness file\n",
+    "res = await ezkl.gen_witness()\n"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "res = ezkl.mock()\n",
+    "assert res == True"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "\n",
+    "# HERE WE SETUP THE CIRCUIT PARAMS\n",
+    "# WE GOT KEYS\n",
+    "# WE GOT CIRCUIT PARAMETERS\n",
+    "# EVERYTHING ANYONE HAS EVER NEEDED FOR ZK\n",
+    "\n",
+    "res = ezkl.setup()\n",
+    "\n",
+    "assert res == True"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "# GENERATE A PROOF\n",
+    "res = ezkl.prove(proof_path=\"proof.json\")\n",
+    "\n",
+    "print(res)\n"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "# VERIFY IT\n",
+    "res = ezkl.verify()\n",
+    "\n",
+    "assert res == True\n",
+    "print(\"verified\")\n",
+    "\n"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "We can also verify it on chain by creating an onchain verifier"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "# check if notebook is in colab\n",
+    "try:\n",
+    "    import google.colab\n",
+    "    import subprocess\n",
+    "    import sys\n",
+    "    subprocess.check_call([sys.executable, \"-m\", \"pip\", \"install\", \"solc-select\"])\n",
+    "    !solc-select install 0.8.20\n",
+    "    !solc-select use 0.8.20\n",
+    "    !solc --version\n",
+    "    import os\n",
+    "\n",
+    "# rely on local installation if the notebook is not in colab\n",
+    "except:\n",
+    "    import os\n",
+    "    pass"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "res = await ezkl.create_evm_verifier()\n",
+    "assert res == True\n"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "You should see a `Verifier.sol`. Right-click and save it locally.\n",
+    "\n",
+    "Now go to [https://remix.ethereum.org](https://remix.ethereum.org).\n",
+    "\n",
+    "Create a new file within remix and copy the verifier code over.\n",
+    "\n",
+    "Finally, compile the code and deploy. For the demo you can deploy to the test environment within remix.\n",
+    "\n",
+    "If everything works, you would have deployed your verifer onchain! Copy the values in the cell above to the respective fields to test if the verifier is working."
+   ]
+  }
+ ],
+ "metadata": {
+  "kernelspec": {
+   "display_name": ".env",
+   "language": "python",
+   "name": "python3"
+  },
+  "language_info": {
+   "codemirror_mode": {
+    "name": "ipython",
+    "version": 3
+   },
+   "file_extension": ".py",
+   "mimetype": "text/x-python",
+   "name": "python",
+   "nbconvert_exporter": "python",
+   "pygments_lexer": "ipython3",
+   "version": "3.12.7"
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 2
+}

examples/notebooks/proof_splitting.ipynb

@@ -232,7 +232,7 @@
                 "run_args.param_visibility = \"fixed\"\n",
                 "run_args.output_visibility = \"public\"\n",
                 "run_args.input_scale = 2\n",
-                "run_args.logrows = 8\n",
+                "run_args.logrows = 15\n",
                 "\n",
                 "ezkl.get_srs(logrows=run_args.logrows, commitment=ezkl.PyCommitments.KZG)"
             ]
@@ -404,7 +404,7 @@
                 "run_args.output_visibility = \"polycommit\"\n",
                 "run_args.variables = [(\"batch_size\", 1)]\n",
                 "run_args.input_scale = 2\n",
-                "run_args.logrows = 8\n"
+                "run_args.logrows = 15\n"
             ]
         },
         {
@@ -466,7 +466,7 @@
             "name": "python",
             "nbconvert_exporter": "python",
             "pygments_lexer": "ipython3",
-            "version": "3.12.2"
+            "version": "3.12.5"
         },
         "orig_nbformat": 4
     },

examples/notebooks/random_forest.ipynb

@@ -152,9 +152,11 @@
             "metadata": {},
             "outputs": [],
             "source": [
-                "!RUST_LOG=trace\n",
-                "# TODO: Dictionary outputs\n",
-                "res = ezkl.gen_settings(model_path, settings_path)\n",
+                "run_args = ezkl.PyRunArgs()\n",
+                "# logrows\n",
+                "run_args.logrows = 20\n",
+                "\n",
+                "res = ezkl.gen_settings(model_path, settings_path, py_run_args=run_args)\n",
                 "assert res == True\n"
             ]
         },
@@ -302,7 +304,7 @@
             "name": "python",
             "nbconvert_exporter": "python",
             "pygments_lexer": "ipython3",
-            "version": "3.12.2"
+            "version": "3.9.13"
         }
     },
     "nbformat": 4,

examples/notebooks/reusable_verifier.ipynb

@@ -0,0 +1,339 @@
+{
+ "cells": [
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "# Reusable Verifiers \n",
+    "\n",
+    "This notebook demonstrates how to create and reuse the same set of separated verifiers for different models. Specifically, we will use the same verifier for the following four models:\n",
+    "\n",
+    "- `1l_mlp sigmoid`\n",
+    "- `1l_mlp relu`\n",
+    "- `1l_conv sigmoid`\n",
+    "- `1l_conv relu`\n",
+    "\n",
+    "When deploying EZKL verifiers on the blockchain, each associated model typically requires its own unique verifier, leading to increased on-chain state usage. \n",
+    "However, with the reusable verifier, we can deploy a single verifier that can be used to verify proofs for any valid H2 circuit. This notebook shows how to do so. \n",
+    "\n",
+    "By reusing the same verifier across multiple models, we significantly reduce the amount of state bloat on the blockchain. Instead of deploying a unique verifier for each model, we deploy a unique and much smaller verifying key artifact (VKA) contract for each model while sharing a common separated verifier. The VKA contains the VK for the model as well circuit specific metadata that was otherwise hardcoded into the stack of the original non-reusable verifier."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "import torch\n",
+    "import torch.nn as nn\n",
+    "import torch.onnx\n",
+    "\n",
+    "# Define the models\n",
+    "class MLP_Sigmoid(nn.Module):\n",
+    "    def __init__(self):\n",
+    "        super(MLP_Sigmoid, self).__init__()\n",
+    "        self.fc = nn.Linear(3, 3)\n",
+    "        self.sigmoid = nn.Sigmoid()\n",
+    "\n",
+    "    def forward(self, x):\n",
+    "        x = self.fc(x)\n",
+    "        x = self.sigmoid(x)\n",
+    "        return x\n",
+    "\n",
+    "class MLP_Relu(nn.Module):\n",
+    "    def __init__(self):\n",
+    "        super(MLP_Relu, self).__init__()\n",
+    "        self.fc = nn.Linear(3, 3)\n",
+    "        self.relu = nn.ReLU()\n",
+    "\n",
+    "    def forward(self, x):\n",
+    "        x = self.fc(x)\n",
+    "        x = self.relu(x)\n",
+    "        return x\n",
+    "\n",
+    "class Conv_Sigmoid(nn.Module):\n",
+    "    def __init__(self):\n",
+    "        super(Conv_Sigmoid, self).__init__()\n",
+    "        self.conv = nn.Conv1d(1, 1, kernel_size=3, stride=1)\n",
+    "        self.sigmoid = nn.Sigmoid()\n",
+    "\n",
+    "    def forward(self, x):\n",
+    "        x = self.conv(x)\n",
+    "        x = self.sigmoid(x)\n",
+    "        return x\n",
+    "\n",
+    "class Conv_Relu(nn.Module):\n",
+    "    def __init__(self):\n",
+    "        super(Conv_Relu, self).__init__()\n",
+    "        self.conv = nn.Conv1d(1, 1, kernel_size=3, stride=1)\n",
+    "        self.relu = nn.ReLU()\n",
+    "\n",
+    "    def forward(self, x):\n",
+    "        x = self.conv(x)\n",
+    "        x = self.relu(x)\n",
+    "        return x\n",
+    "\n",
+    "# Instantiate the models\n",
+    "mlp_sigmoid = MLP_Sigmoid()\n",
+    "mlp_relu = MLP_Relu()\n",
+    "conv_sigmoid = Conv_Sigmoid()\n",
+    "conv_relu = Conv_Relu()\n",
+    "\n",
+    "# Dummy input tensor for mlp\n",
+    "dummy_input_mlp = torch.tensor([[-1.5737053155899048, -1.708398461341858, 0.19544155895709991]])\n",
+    "input_mlp_path = 'mlp_input.json'\n",
+    "\n",
+    "# Dummy input tensor for conv\n",
+    "dummy_input_conv = torch.tensor([[[1.4124163389205933, 0.6938204169273376, 1.0664031505584717]]])\n",
+    "input_conv_path = 'conv_input.json'"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "names = ['mlp_sigmoid', 'mlp_relu', 'conv_sigmoid', 'conv_relu']\n",
+    "models = [mlp_sigmoid, mlp_relu, conv_sigmoid, conv_relu]\n",
+    "inputs = [dummy_input_mlp, dummy_input_mlp, dummy_input_conv, dummy_input_conv]\n",
+    "input_paths = [input_mlp_path, input_mlp_path, input_conv_path, input_conv_path]"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "import os\n",
+    "import json\n",
+    "import torch\n",
+    "import ezkl\n",
+    "\n",
+    "for name, model, x, input_path in zip(names, models, inputs, input_paths):\n",
+    "    # Create a new directory for the model if it doesn't exist\n",
+    "    if not os.path.exists(name):\n",
+    "        os.mkdir(name)\n",
+    "    # Store the paths in each of their respective directories\n",
+    "    model_path = os.path.join(name, \"network.onnx\")\n",
+    "    compiled_model_path = os.path.join(name, \"network.compiled\")\n",
+    "    pk_path = os.path.join(name, \"test.pk\")\n",
+    "    vk_path = os.path.join(name, \"test.vk\")\n",
+    "    settings_path = os.path.join(name, \"settings.json\")\n",
+    "\n",
+    "    witness_path = os.path.join(name, \"witness.json\")\n",
+    "    sol_code_path = os.path.join(name, 'test.sol')\n",
+    "    sol_key_code_path = os.path.join(name, 'test_key.sol')\n",
+    "    abi_path = os.path.join(name, 'test.abi')\n",
+    "    proof_path = os.path.join(name, \"proof.json\")\n",
+    "\n",
+    "    # Flips the neural net into inference mode\n",
+    "    model.eval()\n",
+    "\n",
+    "    # Export the model\n",
+    "    torch.onnx.export(model, x, model_path, export_params=True, opset_version=10,\n",
+    "                      do_constant_folding=True, input_names=['input'],\n",
+    "                      output_names=['output'], dynamic_axes={'input': {0: 'batch_size'},\n",
+    "                                                              'output': {0: 'batch_size'}})\n",
+    "\n",
+    "    data_array = ((x).detach().numpy()).reshape([-1]).tolist()\n",
+    "    data = dict(input_data=[data_array])\n",
+    "    json.dump(data, open(input_path, 'w'))\n",
+    "\n",
+    "    py_run_args = ezkl.PyRunArgs()\n",
+    "    py_run_args.input_visibility = \"private\"\n",
+    "    py_run_args.output_visibility = \"public\"\n",
+    "    py_run_args.param_visibility = \"fixed\"  # private by default\n",
+    "\n",
+    "    res = ezkl.gen_settings(model_path, settings_path, py_run_args=py_run_args)\n",
+    "    assert res == True\n",
+    "\n",
+    "    await ezkl.calibrate_settings(input_path, model_path, settings_path, \"resources\")\n",
+    "\n",
+    "    res = ezkl.compile_circuit(model_path, compiled_model_path, settings_path)\n",
+    "    assert res == True\n",
+    "\n",
+    "    res = await ezkl.get_srs(settings_path)\n",
+    "    assert res == True\n",
+    "\n",
+    "    # now generate the witness file\n",
+    "    res = await ezkl.gen_witness(input_path, compiled_model_path, witness_path)\n",
+    "    assert os.path.isfile(witness_path) == True\n",
+    "\n",
+    "    # SETUP \n",
+    "    # We recommend disabling selector compression for the setup as it decreases the size of the VK artifact\n",
+    "    res = ezkl.setup(compiled_model_path, vk_path, pk_path, disable_selector_compression=True)\n",
+    "    assert res == True\n",
+    "    assert os.path.isfile(vk_path)\n",
+    "    assert os.path.isfile(pk_path)\n",
+    "    assert os.path.isfile(settings_path)\n",
+    "\n",
+    "    # GENERATE A PROOF\n",
+    "    res = ezkl.prove(witness_path, compiled_model_path, pk_path, proof_path, \"single\")\n",
+    "    assert os.path.isfile(proof_path)\n",
+    "\n",
+    "    res = await ezkl.create_evm_verifier(vk_path, settings_path, sol_code_path, abi_path, reusable=True)\n",
+    "    assert res == True\n",
+    "\n",
+    "    res = await ezkl.create_evm_vka(vk_path, settings_path, sol_key_code_path, abi_path)\n",
+    "    assert res == True\n"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "import subprocess\n",
+    "import time\n",
+    "\n",
+    "# make sure anvil is running locally\n",
+    "# $ anvil -p 3030\n",
+    "\n",
+    "RPC_URL = \"http://localhost:3030\"\n",
+    "\n",
+    "# Save process globally\n",
+    "anvil_process = None\n",
+    "\n",
+    "def start_anvil():\n",
+    "    global anvil_process\n",
+    "    if anvil_process is None:\n",
+    "        anvil_process = subprocess.Popen([\"anvil\", \"-p\", \"3030\", \"--code-size-limit=41943040\"])\n",
+    "        if anvil_process.returncode is not None:\n",
+    "            raise Exception(\"failed to start anvil process\")\n",
+    "        time.sleep(3)\n",
+    "\n",
+    "def stop_anvil():\n",
+    "    global anvil_process\n",
+    "    if anvil_process is not None:\n",
+    "        anvil_process.terminate()\n",
+    "        anvil_process = None\n"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "Check that the generated verifiers are identical for all models."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "start_anvil()"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "import filecmp\n",
+    "\n",
+    "def compare_files(file1, file2):\n",
+    "    return filecmp.cmp(file1, file2, shallow=False)\n",
+    "\n",
+    "sol_code_path_0 = os.path.join(\"mlp_sigmoid\", 'test.sol')\n",
+    "sol_code_path_1 = os.path.join(\"mlp_relu\", 'test.sol')\n",
+    "\n",
+    "sol_code_path_2 = os.path.join(\"conv_sigmoid\", 'test.sol')\n",
+    "sol_code_path_3 = os.path.join(\"conv_relu\", 'test.sol')\n",
+    "\n",
+    "\n",
+    "assert compare_files(sol_code_path_0, sol_code_path_1) == True\n",
+    "assert compare_files(sol_code_path_2, sol_code_path_3) == True"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "Here we deploy separate verifier that will be shared by the four models. We picked the `1l_mlp sigmoid` model as an example but you could have used any of the generated verifiers since they are all identical. "
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "import os \n",
+    "addr_path_verifier = \"addr_verifier.txt\"\n",
+    "sol_code_path = os.path.join(\"mlp_sigmoid\", 'test.sol')\n",
+    "\n",
+    "res = await ezkl.deploy_evm(\n",
+    "    addr_path_verifier,\n",
+    "    sol_code_path,\n",
+    "    'http://127.0.0.1:3030',\n",
+    "    \"verifier/reusable\"\n",
+    ")\n",
+    "\n",
+    "assert res == True\n",
+    "\n",
+    "with open(addr_path_verifier, 'r') as file:\n",
+    "    addr = file.read().rstrip()"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "Finally we deploy each of the unique VK-artifacts and verify them using the shared verifier deployed in the previous step."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "for name in names:\n",
+    "    addr_path_vk = \"addr_vk.txt\"\n",
+    "    sol_key_code_path = os.path.join(name, 'test_key.sol')\n",
+    "    res = await ezkl.deploy_evm(addr_path_vk, sol_key_code_path, 'http://127.0.0.1:3030', \"vka\")\n",
+    "    assert res == True\n",
+    "\n",
+    "    with open(addr_path_vk, 'r') as file:\n",
+    "        addr_vk = file.read().rstrip()\n",
+    "        \n",
+    "    proof_path = os.path.join(name, \"proof.json\")\n",
+    "    sol_code_path = os.path.join(name, 'vk.sol')\n",
+    "    res = await ezkl.verify_evm(\n",
+    "        addr,\n",
+    "        proof_path,\n",
+    "        \"http://127.0.0.1:3030\",\n",
+    "        addr_vk = addr_vk\n",
+    "    )\n",
+    "    assert res == True"
+   ]
+  }
+ ],
+ "metadata": {
+  "kernelspec": {
+   "display_name": ".env",
+   "language": "python",
+   "name": "python3"
+  },
+  "language_info": {
+   "codemirror_mode": {
+    "name": "ipython",
+    "version": 3
+   },
+   "file_extension": ".py",
+   "mimetype": "text/x-python",
+   "name": "python",
+   "nbconvert_exporter": "python",
+   "pygments_lexer": "ipython3",
+   "version": "3.11.5"
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 2
+}

examples/notebooks/set_membership.ipynb

@@ -167,6 +167,8 @@
         "run_args = ezkl.PyRunArgs()\n",
         "# \"hashed/private\" means that the output of the hashing is not visible to the verifier and is instead fed into the computational graph\n",
         "run_args.input_visibility = \"hashed/private/0\"\n",
+        "# as the inputs are felts we turn off input range checks\n",
+        "run_args.ignore_range_check_inputs_outputs = True\n",
         "# we set it to fix the set we want to check membership for\n",
         "run_args.param_visibility = \"fixed\"\n",
         "# the output is public -- set membership fails if it is not = 0\n",
@@ -519,4 +521,4 @@
   },
   "nbformat": 4,
   "nbformat_minor": 5
-}
+}

examples/notebooks/simple_demo_aggregated_proofs.ipynb

@@ -171,7 +171,7 @@
                 "json.dump(data, open(cal_path, 'w'))\n",
                 "\n",
                 "\n",
-                "ezkl.calibrate_settings(cal_path, model_path, settings_path, \"resources\")"
+                "await ezkl.calibrate_settings(cal_path, model_path, settings_path, \"resources\")"
             ]
         },
         {
@@ -328,7 +328,7 @@
         },
         {
             "cell_type": "code",
-            "execution_count": null,
+            "execution_count": 26,
             "id": "171702d3",
             "metadata": {},
             "outputs": [],
@@ -348,7 +348,7 @@
         },
         {
             "cell_type": "code",
-            "execution_count": null,
+            "execution_count": 27,
             "id": "671dfdd5",
             "metadata": {},
             "outputs": [],
@@ -364,7 +364,7 @@
         },
         {
             "cell_type": "code",
-            "execution_count": null,
+            "execution_count": 28,
             "id": "50eba2f4",
             "metadata": {},
             "outputs": [],
@@ -399,9 +399,9 @@
             "name": "python",
             "nbconvert_exporter": "python",
             "pygments_lexer": "ipython3",
-            "version": "3.9.15"
+            "version": "3.12.7"
         }
     },
     "nbformat": 4,
     "nbformat_minor": 5
-}
+}

examples/notebooks/solvency.ipynb

@@ -204,6 +204,7 @@
         "run_args = ezkl.PyRunArgs()\n",
         "# \"polycommit\" means that the output of the hashing is not visible to the verifier and is instead fed into the computational graph\n",
         "run_args.input_visibility = \"polycommit\"\n",
+        "run_args.ignore_range_check_inputs_outputs = True\n",
         "# the parameters are public\n",
         "run_args.param_visibility = \"fixed\"\n",
         "# the output is public (this is the inequality test)\n",
@@ -514,4 +515,4 @@
   },
   "nbformat": 4,
   "nbformat_minor": 5
-}
+}

examples/notebooks/univ3-da.ipynb

@@ -0,0 +1,764 @@
+{
+    "cells": [
+        {
+            "attachments": {},
+            "cell_type": "markdown",
+            "metadata": {},
+            "source": [
+                "# univ3-da-ezkl\n",
+                "\n",
+                "Here's an example leveraging EZKL whereby the inputs to the model are read and attested to from an on-chain source. For this setup we make a single call to a view function that returns an array of UniV3 historical TWAP price data that we will attest to on-chain. \n"
+            ]
+        },
+        {
+            "attachments": {},
+            "cell_type": "markdown",
+            "metadata": {},
+            "source": [
+                "First we import the necessary dependencies and set up logging to be as informative as possible. "
+            ]
+        },
+        {
+            "cell_type": "code",
+            "execution_count": null,
+            "metadata": {},
+            "outputs": [],
+            "source": [
+                "# check if notebook is in colab\n",
+                "try:\n",
+                "    # install ezkl\n",
+                "    import google.colab\n",
+                "    import subprocess\n",
+                "    import sys\n",
+                "    subprocess.check_call([sys.executable, \"-m\", \"pip\", \"install\", \"ezkl\"])\n",
+                "    subprocess.check_call([sys.executable, \"-m\", \"pip\", \"install\", \"onnx\"])\n",
+                "\n",
+                "# rely on local installation of ezkl if the notebook is not in colab\n",
+                "except:\n",
+                "    pass\n",
+                "\n",
+                "\n",
+                "from torch import nn\n",
+                "import ezkl\n",
+                "import os\n",
+                "import json\n",
+                "import logging\n",
+                "\n",
+                "# uncomment for more descriptive logging \n",
+                "FORMAT = '%(levelname)s %(name)s %(asctime)-15s %(filename)s:%(lineno)d %(message)s'\n",
+                "logging.basicConfig(format=FORMAT)\n",
+                "logging.getLogger().setLevel(logging.DEBUG)\n"
+            ]
+        },
+        {
+            "attachments": {},
+            "cell_type": "markdown",
+            "metadata": {},
+            "source": [
+                "Now we define our model. It is a very simple PyTorch model that has just one layer, an average pooling 2D layer. "
+            ]
+        },
+        {
+            "cell_type": "code",
+            "execution_count": null,
+            "metadata": {},
+            "outputs": [],
+            "source": [
+                "import torch\n",
+                "# Defines the model\n",
+                "\n",
+                "class MyModel(nn.Module):\n",
+                "    def __init__(self):\n",
+                "        super(MyModel, self).__init__()\n",
+                "        self.layer = nn.AvgPool2d(2, 1, (1, 1))\n",
+                "\n",
+                "    def forward(self, x):\n",
+                "        return self.layer(x)[0]\n",
+                "\n",
+                "\n",
+                "circuit = MyModel()\n",
+                "\n",
+                "# this is where you'd train your model"
+            ]
+        },
+        {
+            "attachments": {},
+            "cell_type": "markdown",
+            "metadata": {},
+            "source": [
+                "We omit training for purposes of this demonstration. We've marked where training would happen in the cell above. \n",
+                "Now we export the model to onnx and create a corresponding (randomly generated) input. This input data will eventually be stored on chain and read from according to the call_data field in the graph input.\n",
+                "\n",
+                "You can replace the random `x` with real data if you so wish. "
+            ]
+        },
+        {
+            "cell_type": "code",
+            "execution_count": null,
+            "metadata": {},
+            "outputs": [],
+            "source": [
+                "x = 0.1*torch.rand(1,*[3, 2, 2], requires_grad=True)\n",
+                "\n",
+                "# Flips the neural net into inference mode\n",
+                "circuit.eval()\n",
+                "\n",
+                "    # Export the model\n",
+                "torch.onnx.export(circuit,               # model being run\n",
+                "                      x,                   # model input (or a tuple for multiple inputs)\n",
+                "                      \"network.onnx\",            # where to save the model (can be a file or file-like object)\n",
+                "                      export_params=True,        # store the trained parameter weights inside the model file\n",
+                "                      opset_version=10,          # the ONNX version to export the model to\n",
+                "                      do_constant_folding=True,  # whether to execute constant folding for optimization\n",
+                "                      input_names = ['input'],   # the model's input names\n",
+                "                      output_names = ['output'], # the model's output names\n",
+                "                      dynamic_axes={'input' : {0 : 'batch_size'},    # variable length axes\n",
+                "                                    'output' : {0 : 'batch_size'}})\n",
+                "\n",
+                "data_array = ((x).detach().numpy()).reshape([-1]).tolist()\n",
+                "\n",
+                "data = dict(input_data = [data_array])\n",
+                "\n",
+                "    # Serialize data into file:\n",
+                "json.dump(data, open(\"input.json\", 'w' ))\n",
+                "\n"
+            ]
+        },
+        {
+            "attachments": {},
+            "cell_type": "markdown",
+            "metadata": {},
+            "source": [
+                "We now define a function that will create a new anvil instance which we will deploy our test contract too. This contract will contain in its storage the data that we will read from and attest to. In production you would not need to set up a local anvil instance. Instead you would replace RPC_URL with the actual RPC endpoint of the chain you are deploying your verifiers too, reading from the data on said chain."
+            ]
+        },
+        {
+            "cell_type": "code",
+            "execution_count": null,
+            "metadata": {},
+            "outputs": [],
+            "source": [
+                "import subprocess\n",
+                "import time\n",
+                "import threading\n",
+                "\n",
+                "# make sure anvil is running locally\n",
+                "# $ anvil -p 3030\n",
+                "\n",
+                "RPC_URL = \"http://localhost:3030\"\n",
+                "\n",
+                "# Save process globally\n",
+                "anvil_process = None\n",
+                "\n",
+                "def start_anvil():\n",
+                "    global anvil_process\n",
+                "    if anvil_process is None:\n",
+                "        anvil_process = subprocess.Popen([\"anvil\", \"-p\", \"3030\", \"--fork-url\", \"https://arb1.arbitrum.io/rpc\", \"--code-size-limit=41943040\"])\n",
+                "        if anvil_process.returncode is not None:\n",
+                "            raise Exception(\"failed to start anvil process\")\n",
+                "        time.sleep(3)\n",
+                "\n",
+                "def stop_anvil():\n",
+                "    global anvil_process\n",
+                "    if anvil_process is not None:\n",
+                "        anvil_process.terminate()\n",
+                "        anvil_process = None\n"
+            ]
+        },
+        {
+            "attachments": {},
+            "cell_type": "markdown",
+            "metadata": {},
+            "source": [
+                "We define our `PyRunArgs` objects which contains the visibility parameters for out model. \n",
+                "- `input_visibility` defines the visibility of the model inputs\n",
+                "- `param_visibility` defines the visibility of the model weights and constants and parameters \n",
+                "- `output_visibility` defines the visibility of the model outputs\n",
+                "\n",
+                "Here we create the following setup:\n",
+                "- `input_visibility`: \"public\"\n",
+                "- `param_visibility`: \"private\"\n",
+                "- `output_visibility`: public\n"
+            ]
+        },
+        {
+            "cell_type": "code",
+            "execution_count": null,
+            "metadata": {},
+            "outputs": [],
+            "source": [
+                "import ezkl\n",
+                "\n",
+                "model_path = os.path.join('network.onnx')\n",
+                "compiled_model_path = os.path.join('network.compiled')\n",
+                "pk_path = os.path.join('test.pk')\n",
+                "vk_path = os.path.join('test.vk')\n",
+                "settings_path = os.path.join('settings.json')\n",
+                "srs_path = os.path.join('kzg.srs')\n",
+                "data_path = os.path.join('input.json')\n",
+                "\n",
+                "run_args = ezkl.PyRunArgs()\n",
+                "run_args.input_visibility = \"public\"\n",
+                "run_args.param_visibility = \"private\"\n",
+                "run_args.output_visibility = \"public\"\n",
+                "run_args.decomp_legs=6\n",
+                "run_args.num_inner_cols = 1\n",
+                "run_args.variables = [(\"batch_size\", 1)]"
+            ]
+        },
+        {
+            "attachments": {},
+            "cell_type": "markdown",
+            "metadata": {},
+            "source": [
+                "Now we generate a settings file. This file basically instantiates a bunch of parameters that determine their circuit shape, size etc... Because of the way we represent nonlinearities in the circuit (using Halo2's [lookup tables](https://zcash.github.io/halo2/design/proving-system/lookup.html)), it is often best to _calibrate_ this settings file as some data can fall out of range of these lookups.\n",
+                "\n",
+                "You can pass a dataset for calibration that will be representative of real inputs you might find if and when you deploy the prover. Here we create a dummy calibration dataset for demonstration purposes. "
+            ]
+        },
+        {
+            "cell_type": "code",
+            "execution_count": null,
+            "metadata": {},
+            "outputs": [],
+            "source": [
+                "# TODO: Dictionary outputs\n",
+                "res = ezkl.gen_settings(model_path, settings_path, py_run_args=run_args)\n",
+                "assert res == True"
+            ]
+        },
+        {
+            "cell_type": "code",
+            "execution_count": null,
+            "metadata": {},
+            "outputs": [],
+            "source": [
+                "# generate a bunch of dummy calibration data\n",
+                "cal_data = {\n",
+                "    \"input_data\": [(0.1*torch.rand(2, *[3, 2, 2])).flatten().tolist()],\n",
+                "}\n",
+                "\n",
+                "cal_path = os.path.join('val_data.json')\n",
+                "# save as json file\n",
+                "with open(cal_path, \"w\") as f:\n",
+                "    json.dump(cal_data, f)\n",
+                "\n",
+                "res = await ezkl.calibrate_settings(cal_path, model_path, settings_path, \"resources\")"
+            ]
+        },
+        {
+            "cell_type": "code",
+            "execution_count": null,
+            "metadata": {},
+            "outputs": [],
+            "source": [
+                "res = ezkl.compile_circuit(model_path, compiled_model_path, settings_path)\n",
+                "assert res == True"
+            ]
+        },
+        {
+            "cell_type": "markdown",
+            "metadata": {},
+            "source": [
+                "The graph input for on chain data sources is formatted completely differently compared to file based data sources.\n",
+                "\n",
+                "- For file data sources, the raw floating point values that eventually get quantized, converted into field elements and stored in `witness.json` to be consumed by the circuit are stored. The output data contains the expected floating point values returned as outputs from running your vanilla pytorch model on the given inputs.\n",
+                "- For on chain data sources, the input_data field contains all the data necessary to read and format the on chain data into something digestable by EZKL (aka field elements :-D). \n",
+                "Here is what the schema for an on-chain data source graph input file should look like for a single call data source:\n",
+                "    \n",
+                "```json\n",
+                "{\n",
+                "    \"input_data\": {\n",
+                "        \"rpc\": \"http://localhost:3030\", // The rpc endpoint of the chain you are deploying your verifier to\n",
+                "        \"calls\": {\n",
+                "            \"call_data\": \"1f3be514000000000000000000000000c6962004f452be9203591991d15f6b388e09e8d00000000000000000000000000000000000000000000000000000000000000040000000000000000000000000000000000000000000000000000000000000000c000000000000000000000000000000000000000000000000000000000000000b000000000000000000000000000000000000000000000000000000000000000a0000000000000000000000000000000000000000000000000000000000000009000000000000000000000000000000000000000000000000000000000000000800000000000000000000000000000000000000000000000000000000000000070000000000000000000000000000000000000000000000000000000000000006000000000000000000000000000000000000000000000000000000000000000500000000000000000000000000000000000000000000000000000000000000040000000000000000000000000000000000000000000000000000000000000003000000000000000000000000000000000000000000000000000000000000000200000000000000000000000000000000000000000000000000000000000000010000000000000000000000000000000000000000000000000000000000000000\", // The abi encoded call data to a view function that returns an array of on-chain data points we are attesting to. \n",
+                "            \"decimals\": 0, // The number of decimal places of the large uint256 value. This is our way of representing large wei values as floating points on chain, since the evm only natively supports integer values.\n",
+                "            \"address\": \"9A213F53334279C128C37DA962E5472eCD90554f\", // The address of the contract that we are calling to get the data. \n",
+                "            \"len\": 12 // The number of data points returned by the view function (the length of the array)\n",
+                "        }\n",
+                "    }\n",
+                "}\n",
+                "```"
+            ]
+        },
+        {
+            "cell_type": "code",
+            "execution_count": null,
+            "metadata": {},
+            "outputs": [],
+            "source": [
+                "from web3 import Web3, HTTPProvider\n",
+                "from solcx import compile_standard\n",
+                "from decimal import Decimal\n",
+                "import json\n",
+                "import os\n",
+                "import torch\n",
+                "import requests\n",
+                "\n",
+                "# This function counts the decimal places of a floating point number\n",
+                "def count_decimal_places(num):\n",
+                "    num_str = str(num)\n",
+                "    if '.' in num_str:\n",
+                "        return len(num_str) - 1 - num_str.index('.')\n",
+                "    else:\n",
+                "        return 0\n",
+                "\n",
+                "# setup web3 instance\n",
+                "w3 = Web3(HTTPProvider(RPC_URL)) \n",
+                "\n",
+                "def set_next_block_timestamp(anvil_url, timestamp):\n",
+                "    # Send the JSON-RPC request to Anvil\n",
+                "    payload = {\n",
+                "        \"jsonrpc\": \"2.0\",\n",
+                "        \"id\": 1,\n",
+                "        \"method\": \"evm_setNextBlockTimestamp\",\n",
+                "        \"params\": [timestamp]\n",
+                "    }\n",
+                "    response = requests.post(anvil_url, json=payload)\n",
+                "    if response.status_code == 200:\n",
+                "        print(f\"Next block timestamp set to: {timestamp}\")\n",
+                "    else:\n",
+                "        print(f\"Failed to set next block timestamp: {response.text}\")\n",
+                "\n",
+                "def on_chain_data(tensor):\n",
+                "    # Step 0: Convert the tensor to a flat list\n",
+                "    data = tensor.view(-1).tolist()\n",
+                "\n",
+                "    # Step 1: Prepare the calldata\n",
+                "    secondsAgo = [len(data) - 1 - i for i in range(len(data))]\n",
+                "\n",
+                "    # Step 2: Prepare and compile the contract UniTickAttestor contract\n",
+                "    contract_source_code = '''\n",
+                "    // SPDX-License-Identifier: MIT\n",
+                "    pragma solidity ^0.8.20;\n",
+                "\n",
+                "    /// @title Pool state that is not stored\n",
+                "    /// @notice Contains view functions to provide information about the pool that is computed rather than stored on the\n",
+                "    /// blockchain. The functions here may have variable gas costs.\n",
+                "    interface IUniswapV3PoolDerivedState {\n",
+                "        /// @notice Returns the cumulative tick and liquidity as of each timestamp `secondsAgo` from the current block timestamp\n",
+                "        /// @dev To get a time weighted average tick or liquidity-in-range, you must call this with two values, one representing\n",
+                "        /// the beginning of the period and another for the end of the period. E.g., to get the last hour time-weighted average tick,\n",
+                "        /// you must call it with secondsAgos = [3600, 0].\n",
+                "        /// log base sqrt(1.0001) of token1 / token0. The TickMath library can be used to go from a tick value to a ratio.\n",
+                "        /// @dev The time weighted average tick represents the geometric time weighted average price of the pool, in\n",
+                "        /// @param secondsAgos From how long ago each cumulative tick and liquidity value should be returned\n",
+                "        /// @return tickCumulatives Cumulative tick values as of each `secondsAgos` from the current block timestamp\n",
+                "        /// @return secondsPerLiquidityCumulativeX128s Cumulative seconds per liquidity-in-range value as of each `secondsAgos` from the current block\n",
+                "        /// timestamp\n",
+                "        function observe(\n",
+                "            uint32[] calldata secondsAgos\n",
+                "        )\n",
+                "            external\n",
+                "            view\n",
+                "            returns (\n",
+                "                int56[] memory tickCumulatives,\n",
+                "                uint160[] memory secondsPerLiquidityCumulativeX128s\n",
+                "            );\n",
+                "    }\n",
+                "\n",
+                "    /// @title Uniswap Wrapper around `pool.observe` that stores the parameters for fetching and then attesting to historical data\n",
+                "    /// @notice Provides functions to integrate with V3 pool oracle\n",
+                "    contract UniTickAttestor {\n",
+                "        /**\n",
+                "        * @notice Calculates time-weighted means of tick and liquidity for a given Uniswap V3 pool\n",
+                "        * @param pool Address of the pool that we want to observe\n",
+                "        * @param secondsAgo Number of seconds in the past from which to calculate the time-weighted means\n",
+                "        * @return tickCumulatives The cumulative tick values as of each `secondsAgo` from the current block timestamp\n",
+                "        */\n",
+                "        function consult(\n",
+                "            IUniswapV3PoolDerivedState pool,\n",
+                "            uint32[] memory secondsAgo\n",
+                "        ) public view returns (int256[] memory tickCumulatives) {\n",
+                "            tickCumulatives = new int256[](secondsAgo.length);\n",
+                "            (int56[] memory _ticks,) = pool.observe(secondsAgo);\n",
+                "            for (uint256 i = 0; i < secondsAgo.length; i++) {\n",
+                "                tickCumulatives[i] = int256(_ticks[i]);\n",
+                "            }\n",
+                "        }\n",
+                "    }\n",
+                "    '''\n",
+                "\n",
+                "    compiled_sol = compile_standard({\n",
+                "        \"language\": \"Solidity\",\n",
+                "        \"sources\": {\"UniTickAttestor.sol\": {\"content\": contract_source_code}},\n",
+                "        \"settings\": {\"outputSelection\": {\"*\": {\"*\": [\"metadata\", \"evm.bytecode\", \"abi\"]}}}\n",
+                "    })\n",
+                "\n",
+                "    # Get bytecode\n",
+                "    bytecode = compiled_sol['contracts']['UniTickAttestor.sol']['UniTickAttestor']['evm']['bytecode']['object']\n",
+                "\n",
+                "    # Get ABI\n",
+                "    # In production if you are reading from really large contracts you can just use\n",
+                "    # a stripped down version of the ABI of the contract you are calling, containing only the view functions you will fetch data from.\n",
+                "    abi = json.loads(compiled_sol['contracts']['UniTickAttestor.sol']['UniTickAttestor']['metadata'])['output']['abi']\n",
+                "\n",
+                "    # Step 3: Deploy the contract\n",
+                "    UniTickAttestor = w3.eth.contract(abi=abi, bytecode=bytecode)\n",
+                "    tx_hash = UniTickAttestor.constructor().transact()\n",
+                "    tx_receipt = w3.eth.wait_for_transaction_receipt(tx_hash)\n",
+                "    # If you are deploying to production you can skip the 3 lines of code above and just instantiate the contract like this,\n",
+                "    # passing the address and abi of the contract you are fetching data from.\n",
+                "    contract = w3.eth.contract(address=tx_receipt['contractAddress'], abi=abi)\n",
+                "\n",
+                "    # Step 4: Interact with the contract\n",
+                "    call = contract.functions.consult(\n",
+                "        # Address of the UniV3 usdc-weth pool 0.005 fee\n",
+                "        \"0xC6962004f452bE9203591991D15f6b388e09E8D0\",\n",
+                "        secondsAgo\n",
+                "    ).build_transaction()\n",
+                "    result = contract.functions.consult(\n",
+                "        # Address of the UniV3 usdc-weth pool 0.005 fee\n",
+                "        \"0xC6962004f452bE9203591991D15f6b388e09E8D0\",\n",
+                "        secondsAgo\n",
+                "    ).call()\n",
+                "    \n",
+                "    print(f'result: {result}')\n",
+                "    calldata = call['data'][2:]\n",
+                "\n",
+                "    time_stamp = w3.eth.get_block('latest')['timestamp']\n",
+                "\n",
+                "    print(f'time_stamp: {time_stamp}')\n",
+                "\n",
+                "    # Set the next block timestamp using the fetched time_stamp\n",
+                "    set_next_block_timestamp(RPC_URL, time_stamp)\n",
+                "\n",
+                "\n",
+                "    # Prepare the calls_to_account object\n",
+                "    # If you were calling view functions across multiple contracts,\n",
+                "    # you would have multiple entries in the calls_to_account array,\n",
+                "    # one for each contract.\n",
+                "    call_to_account = {\n",
+                "        'call_data': calldata,\n",
+                "        'decimals': 0,\n",
+                "        'address': contract.address[2:], # remove the '0x' prefix\n",
+                "        'len': len(data),\n",
+                "    }\n",
+                "\n",
+                "    print(f'call_to_account: {call_to_account}')\n",
+                "\n",
+                "    return call_to_account\n",
+                "\n",
+                "# Now let's start the Anvil process. You don't need to do this if you are deploying to a non-local chain.\n",
+                "start_anvil()\n",
+                "\n",
+                "# Now let's call our function, passing in the same input tensor we used to export the model 2 cells above.\n",
+                "calls_to_account = on_chain_data(x)\n",
+                "\n",
+                "data = dict(input_data =  {'rpc': RPC_URL, 'calls': calls_to_account })\n",
+                "\n",
+                "# Serialize on-chain data into file:\n",
+                "json.dump(data, open(\"input.json\", 'w'))"
+            ]
+        },
+        {
+            "attachments": {},
+            "cell_type": "markdown",
+            "metadata": {},
+            "source": [
+                "As we use Halo2 with KZG-commitments we need an SRS string from (preferably) a multi-party trusted setup ceremony. For an overview of the procedures for such a ceremony check out [this page](https://blog.ethereum.org/2023/01/16/announcing-kzg-ceremony). The `get_srs` command retrieves a correctly sized SRS given the calibrated settings file from [here](https://github.com/han0110/halo2-kzg-srs). \n",
+                "\n",
+                "These SRS were generated with [this](https://github.com/privacy-scaling-explorations/perpetualpowersoftau) ceremony. "
+            ]
+        },
+        {
+            "cell_type": "code",
+            "execution_count": null,
+            "metadata": {},
+            "outputs": [],
+            "source": [
+                "res = await ezkl.get_srs( settings_path)"
+            ]
+        },
+        {
+            "attachments": {},
+            "cell_type": "markdown",
+            "metadata": {},
+            "source": [
+                "We now need to generate the circuit witness. These are the model outputs (and any hashes) that are generated when feeding the previously generated `input.json` through the circuit / model. "
+            ]
+        },
+        {
+            "cell_type": "code",
+            "execution_count": null,
+            "metadata": {},
+            "outputs": [],
+            "source": [
+                "# !export RUST_BACKTRACE=1\n",
+                "\n",
+                "witness_path = \"witness.json\"\n",
+                "\n",
+                "res = await ezkl.gen_witness(data_path, compiled_model_path, witness_path)"
+            ]
+        },
+        {
+            "attachments": {},
+            "cell_type": "markdown",
+            "metadata": {},
+            "source": [
+                "Here we setup verifying and proving keys for the circuit. As the name suggests the proving key is needed for ... proving and the verifying key is needed for ... verifying. "
+            ]
+        },
+        {
+            "cell_type": "code",
+            "execution_count": null,
+            "metadata": {},
+            "outputs": [],
+            "source": [
+                "# HERE WE SETUP THE CIRCUIT PARAMS\n",
+                "# WE GOT KEYS\n",
+                "# WE GOT CIRCUIT PARAMETERS\n",
+                "# EVERYTHING ANYONE HAS EVER NEEDED FOR ZK\n",
+                "res = ezkl.setup(\n",
+                "        compiled_model_path,\n",
+                "        vk_path,\n",
+                "        pk_path,\n",
+                "    )\n",
+                "\n",
+                "assert res == True\n",
+                "assert os.path.isfile(vk_path)\n",
+                "assert os.path.isfile(pk_path)\n",
+                "assert os.path.isfile(settings_path)"
+            ]
+        },
+        {
+            "attachments": {},
+            "cell_type": "markdown",
+            "metadata": {},
+            "source": [
+                "Now we generate a full proof. "
+            ]
+        },
+        {
+            "cell_type": "code",
+            "execution_count": null,
+            "metadata": {},
+            "outputs": [],
+            "source": [
+                "# GENERATE A PROOF\n",
+                "\n",
+                "proof_path = os.path.join('test.pf')\n",
+                "\n",
+                "res = ezkl.prove(\n",
+                "        witness_path,\n",
+                "        compiled_model_path,\n",
+                "        pk_path,\n",
+                "        proof_path,\n",
+                "        \"single\",\n",
+                "    )\n",
+                "\n",
+                "print(res)\n",
+                "assert os.path.isfile(proof_path)"
+            ]
+        },
+        {
+            "attachments": {},
+            "cell_type": "markdown",
+            "metadata": {},
+            "source": [
+                "And verify it as a sanity check. "
+            ]
+        },
+        {
+            "cell_type": "code",
+            "execution_count": null,
+            "metadata": {},
+            "outputs": [],
+            "source": [
+                "# VERIFY IT\n",
+                "\n",
+                "res = ezkl.verify(\n",
+                "        proof_path,\n",
+                "        settings_path,\n",
+                "        vk_path,\n",
+                "    )\n",
+                "\n",
+                "assert res == True\n",
+                "print(\"verified\")"
+            ]
+        },
+        {
+            "cell_type": "markdown",
+            "metadata": {},
+            "source": [
+                "We can now create and then deploy a vanilla evm verifier."
+            ]
+        },
+        {
+            "cell_type": "code",
+            "execution_count": null,
+            "metadata": {},
+            "outputs": [],
+            "source": [
+                "abi_path = 'test.abi'\n",
+                "sol_code_path = 'test.sol'\n",
+                "\n",
+                "res = await ezkl.create_evm_verifier(\n",
+                "        vk_path,\n",
+                "        settings_path,\n",
+                "        sol_code_path,\n",
+                "        abi_path,\n",
+                "    )\n",
+                "assert res == True"
+            ]
+        },
+        {
+            "cell_type": "code",
+            "execution_count": null,
+            "metadata": {},
+            "outputs": [],
+            "source": [
+                "import json\n",
+                "\n",
+                "addr_path_verifier = \"addr_verifier.txt\"\n",
+                "\n",
+                "res = await ezkl.deploy_evm(\n",
+                "    addr_path_verifier,\n",
+                "    sol_code_path,\n",
+                "    'http://127.0.0.1:3030'\n",
+                ")\n",
+                "\n",
+                "assert res == True"
+            ]
+        },
+        {
+            "attachments": {},
+            "cell_type": "markdown",
+            "metadata": {},
+            "source": [
+                "With the vanilla verifier deployed, we can now create the data attestation contract, which will read in the instances from the calldata to the verifier, attest to them, call the verifier and then return the result. "
+            ]
+        },
+        {
+            "cell_type": "code",
+            "execution_count": null,
+            "metadata": {},
+            "outputs": [],
+            "source": [
+                "\n",
+                "abi_path = 'test.abi'\n",
+                "sol_code_path = 'test.sol'\n",
+                "input_path = 'input.json'\n",
+                "\n",
+                "res = await ezkl.create_evm_data_attestation(\n",
+                "        input_path,\n",
+                "        settings_path,\n",
+                "        sol_code_path,\n",
+                "        abi_path,\n",
+                "    )"
+            ]
+        },
+        {
+            "attachments": {},
+            "cell_type": "markdown",
+            "metadata": {},
+            "source": [
+                "Now we can deploy the data attest verifier contract. For security reasons, this binding will only deploy to a local anvil instance, using accounts generated by anvil. \n",
+                "So should only be used for testing purposes."
+            ]
+        },
+        {
+            "cell_type": "code",
+            "execution_count": null,
+            "metadata": {},
+            "outputs": [],
+            "source": [
+                "addr_path_da = \"addr_da.txt\"\n",
+                "\n",
+                "res = await ezkl.deploy_da_evm(\n",
+                "        addr_path_da,\n",
+                "        input_path,\n",
+                "        settings_path,\n",
+                "        sol_code_path,\n",
+                "        RPC_URL,\n",
+                "    )\n"
+            ]
+        },
+        {
+            "attachments": {},
+            "cell_type": "markdown",
+            "metadata": {},
+            "source": [
+                "Here we need to regenerate the witness, prove and then verify all within the same cell. This is because we want to reduce the amount of latency between reading on-chain state and verifying it on-chain. This is because the attest input values read from the oracle are time sensitive (their values are derived from computing on block.timestamp) and can change between the time of reading and the time of verifying.\n",
+                "\n",
+                "Call the view only verify method on the contract to verify the proof. Since it is a view function this is safe to use in production since you don't have to pass your private key."
+            ]
+        },
+        {
+            "cell_type": "code",
+            "execution_count": null,
+            "metadata": {},
+            "outputs": [],
+            "source": [
+                "# !export RUST_BACKTRACE=1\n",
+                "\n",
+                "calls_to_account = on_chain_data(x)\n",
+                "\n",
+                "data = dict(input_data =  {'rpc': RPC_URL, 'calls': calls_to_account })\n",
+                "\n",
+                "# Serialize on-chain data into file:\n",
+                "json.dump(data, open(\"input.json\", 'w'))\n",
+                "\n",
+                "# setup web3 instance\n",
+                "w3 = Web3(HTTPProvider(RPC_URL)) \n",
+                "\n",
+                "time_stamp = w3.eth.get_block('latest')['timestamp']\n",
+                "\n",
+                "print(f'time_stamp: {time_stamp}')\n",
+                "\n",
+                "\n",
+                "witness_path = \"witness.json\"\n",
+                "\n",
+                "res = await ezkl.gen_witness(data_path, compiled_model_path, witness_path)\n",
+                "\n",
+                "res = ezkl.prove(\n",
+                "        witness_path,\n",
+                "        compiled_model_path,\n",
+                "        pk_path,\n",
+                "        proof_path,\n",
+                "        \"single\",\n",
+                "    )\n",
+                "\n",
+                "print(res)\n",
+                "assert os.path.isfile(proof_path)\n",
+                "# read the verifier address\n",
+                "addr_verifier = None\n",
+                "with open(addr_path_verifier, 'r') as f:\n",
+                "    addr = f.read()\n",
+                "#read the data attestation address\n",
+                "addr_da = None\n",
+                "with open(addr_path_da, 'r') as f:\n",
+                "    addr_da = f.read()\n",
+                "\n",
+                "res = await ezkl.verify_evm(\n",
+                "    addr,\n",
+                "    proof_path,\n",
+                "    RPC_URL,\n",
+                "    addr_da,\n",
+                ")"
+            ]
+        }
+    ],
+    "metadata": {
+        "kernelspec": {
+            "display_name": ".env",
+            "language": "python",
+            "name": "python3"
+        },
+        "language_info": {
+            "codemirror_mode": {
+                "name": "ipython",
+                "version": 3
+            },
+            "file_extension": ".py",
+            "mimetype": "text/x-python",
+            "name": "python",
+            "nbconvert_exporter": "python",
+            "pygments_lexer": "ipython3",
+            "version": "3.11.5"
+        },
+        "orig_nbformat": 4
+    },
+    "nbformat": 4,
+    "nbformat_minor": 2
+}

examples/onnx/1d_conv/input.json

@@ -0,0 +1,106 @@
+{
+  "input_data": [
+    [
+      8761,
+      7654,
+      8501,
+      2404,
+      6929,
+      8858,
+      5946,
+      3673,
+      4131,
+      3854,
+      8137,
+      8239,
+      9038,
+      6299,
+      1118,
+      9737,
+      208,
+      7954,
+      3691,
+      610,
+      3468,
+      3314,
+      8658,
+      8366,
+      2850,
+      477,
+      6114,
+      232,
+      4601,
+      7420,
+      5713,
+      2936,
+      6061,
+      2870,
+      8421,
+      177,
+      7107,
+      7382,
+      6115,
+      5487,
+      8502,
+      2559,
+      1875,
+      129,
+      8533,
+      8201,
+      8414,
+      4775,
+      9817,
+      3127,
+      8761,
+      7654,
+      8501,
+      2404,
+      6929,
+      8858,
+      5946,
+      3673,
+      4131,
+      3854,
+      8137,
+      8239,
+      9038,
+      6299,
+      1118,
+      9737,
+      208,
+      7954,
+      3691,
+      610,
+      3468,
+      3314,
+      8658,
+      8366,
+      2850,
+      477,
+      6114,
+      232,
+      4601,
+      7420,
+      5713,
+      2936,
+      6061,
+      2870,
+      8421,
+      177,
+      7107,
+      7382,
+      6115,
+      5487,
+      8502,
+      2559,
+      1875,
+      129,
+      8533,
+      8201,
+      8414,
+      4775,
+      9817,
+      3127
+    ]
+  ]
+}

examples/onnx/1l_conv_transpose/settings.json

@@ -0,0 +1 @@
+{"run_args":{"tolerance":{"val":0.0,"scale":1.0},"input_scale":7,"param_scale":7,"scale_rebase_multiplier":10,"lookup_range":[0,0],"logrows":13,"variables":[["batch_size",1]],"input_visibility":"Private","output_visibility":"Public","param_visibility":"Private"},"num_constraints":5619,"total_const_size":513,"model_instance_shapes":[[1,3,10,10]],"model_output_scales":[14],"model_input_scales":[7],"module_sizes":{"kzg":[],"poseidon":[0,[0]],"elgamal":[0,[0]]},"required_lookups":[],"check_mode":"UNSAFE","version":"0.0.0","num_blinding_factors":null}

examples/onnx/boolean/gen.py

@@ -9,7 +9,9 @@ class MyModel(nn.Module):
         super(MyModel, self).__init__()
 
     def forward(self, w, x, y, z):
-        return [((x & y)) == (x & (y | (z ^ w)))]
+        a = (x & y)
+        b = (y & (z ^ w))
+        return [a & b]
 
 
 circuit = MyModel()

examples/onnx/boolean/input.json

@@ -1 +1 @@
-{"input_data": [[false, true, false], [true, false, false], [true, false, false], [false, false, false]]}
+{"input_data": [[false, true, true], [false, true, true], [true, false, false], [false, true, true]]}

examples/onnx/boolean/network.onnx

@@ -1,21 +1,17 @@
-pytorch1.12.1:«
-+
+pytorch2.2.2:„
+*
 input1
-input2
-onnx::Equal_4And_0"And
+input2
+/And_output_0/And"And
 )
-input3
-input
-onnx::Or_5Xor_1"Xor
+input3
+input
 /Xor_output_0/Xor"Xor
-input2
-
-onnx::Or_5onnx::And_6Or_2"Or
-0
-input1
-onnx::And_6
-onnx::Equal_7And_3"And
-6
+5
+input2
+
+/Xor_output_0/And_1_output_0/And_1"And
+5
 
 /And_output_0
 /And_1_output_0output/And_2"And

examples/onnx/exp/gen.py

@@ -0,0 +1,42 @@
+from torch import nn
+import torch
+import json
+import numpy as np
+
+
+class MyModel(nn.Module):
+    def __init__(self):
+        super(MyModel, self).__init__()
+
+    def forward(self, x):          
+        m = torch.exp(x)
+        
+        return m 
+
+
+circuit = MyModel()
+
+x = torch.empty(1, 8).uniform_(0, 1)
+
+out = circuit(x)
+
+print(out)
+
+torch.onnx.export(circuit, x, "network.onnx",
+                  export_params=True,        # store the trained parameter weights inside the model file
+                  opset_version=17,          # the ONNX version to export the model to
+                  do_constant_folding=True,  # whether to execute constant folding for optimization
+                  input_names=['input'],   # the model's input names
+                  output_names=['output'],  # the model's output names
+                  dynamic_axes={'input': {0: 'batch_size'},  # variable length axes
+                                'output': {0: 'batch_size'}})
+
+
+d1 = ((x).detach().numpy()).reshape([-1]).tolist()
+
+data = dict(
+    input_data=[d1],
+)
+
+# Serialize data into file:
+json.dump(data, open("input.json", 'w'))

examples/onnx/exp/input.json

@@ -0,0 +1 @@
+{"input_data": [[0.5801457762718201, 0.6019012331962585, 0.8695418238639832, 0.17170941829681396, 0.500616729259491, 0.353726327419281, 0.6726185083389282, 0.5936906337738037]]}

examples/onnx/exp/network.onnx

@@ -0,0 +1,14 @@
+pytorch2.2.2:o
+
+inputoutput/Exp"Exp
+main_graphZ!
+input
+

+
+batch_size
+b"
+output
+

+
+batch_size
+B

examples/onnx/general_exp/gen.py

@@ -0,0 +1,41 @@
+from torch import nn
+import torch
+import json
+import numpy as np
+
+
+class MyModel(nn.Module):
+    def __init__(self):
+        super(MyModel, self).__init__()
+
+    def forward(self, x):          
+        m = 10**x
+        return m 
+
+
+circuit = MyModel()
+
+x = torch.empty(1, 8).uniform_(0, 1)
+
+out = circuit(x)
+
+print(out)
+
+torch.onnx.export(circuit, x, "network.onnx",
+                  export_params=True,        # store the trained parameter weights inside the model file
+                  opset_version=17,          # the ONNX version to export the model to
+                  do_constant_folding=True,  # whether to execute constant folding for optimization
+                  input_names=['input'],   # the model's input names
+                  output_names=['output'],  # the model's output names
+                  dynamic_axes={'input': {0: 'batch_size'},  # variable length axes
+                                'output': {0: 'batch_size'}})
+
+
+d1 = ((x).detach().numpy()).reshape([-1]).tolist()
+
+data = dict(
+    input_data=[d1],
+)
+
+# Serialize data into file:
+json.dump(data, open("input.json", 'w'))

examples/onnx/general_exp/input.json

@@ -0,0 +1 @@
+{"input_data": [[0.9837989807128906, 0.026381194591522217, 0.3403851389884949, 0.14531707763671875, 0.24652725458145142, 0.7945117354393005, 0.4076554775238037, 0.23064672946929932]]}

examples/onnx/integer_div/gen.py

@@ -0,0 +1,42 @@
+from torch import nn
+import torch
+import json
+import numpy as np
+
+
+class MyModel(nn.Module):
+    def __init__(self):
+        super(MyModel, self).__init__()
+
+    def forward(self, x):          
+        return x // 3
+
+
+circuit = MyModel()
+
+x = torch.randint(0, 10, (1, 2, 2, 8)) 
+
+out = circuit(x)
+
+print(x)
+print(out)
+print(x/3)
+
+torch.onnx.export(circuit, x, "network.onnx",
+                  export_params=True,        # store the trained parameter weights inside the model file
+                  opset_version=17,          # the ONNX version to export the model to
+                  do_constant_folding=True,  # whether to execute constant folding for optimization
+                  input_names=['input'],   # the model's input names
+                  output_names=['output'],  # the model's output names
+                  dynamic_axes={'input': {0: 'batch_size'},  # variable length axes
+                                'output': {0: 'batch_size'}})
+
+
+d1 = ((x).detach().numpy()).reshape([-1]).tolist()
+
+data = dict(
+    input_data=[d1],
+)
+
+# Serialize data into file:
+json.dump(data, open("input.json", 'w'))

examples/onnx/integer_div/input.json

@@ -0,0 +1 @@
+{"input_data": [[3, 4, 0, 9, 2, 6, 2, 5, 1, 5, 3, 5, 5, 7, 0, 2, 6, 1, 4, 4, 1, 9, 7, 7, 5, 8, 2, 0, 1, 5, 9, 8]]}

examples/onnx/log/gen.py

@@ -0,0 +1,42 @@
+from torch import nn
+import torch
+import json
+import numpy as np
+
+
+class MyModel(nn.Module):
+    def __init__(self):
+        super(MyModel, self).__init__()
+
+    def forward(self, x):          
+        m = torch.log(x)
+        
+        return m 
+
+
+circuit = MyModel()
+
+x = torch.empty(1, 8).uniform_(0, 3)
+
+out = circuit(x)
+
+print(out)
+
+torch.onnx.export(circuit, x, "network.onnx",
+                  export_params=True,        # store the trained parameter weights inside the model file
+                  opset_version=17,          # the ONNX version to export the model to
+                  do_constant_folding=True,  # whether to execute constant folding for optimization
+                  input_names=['input'],   # the model's input names
+                  output_names=['output'],  # the model's output names
+                  dynamic_axes={'input': {0: 'batch_size'},  # variable length axes
+                                'output': {0: 'batch_size'}})
+
+
+d1 = ((x).detach().numpy()).reshape([-1]).tolist()
+
+data = dict(
+    input_data=[d1],
+)
+
+# Serialize data into file:
+json.dump(data, open("input.json", 'w'))

examples/onnx/log/input.json

@@ -0,0 +1 @@
+{"input_data": [[1.9252371788024902, 1.8418371677398682, 0.8400403261184692, 2.083845853805542, 0.9760497808456421, 0.6940176486968994, 0.015579521656036377, 2.2689192295074463]]}

examples/onnx/log/network.onnx

@@ -0,0 +1,14 @@
+pytorch2.2.2:o
+
+inputoutput/Log"Log
+main_graphZ!
+input
+

+
+batch_size
+b"
+output
+

+
+batch_size
+B

examples/onnx/rounding_ops/gen.py

@@ -21,9 +21,9 @@ def main():
     torch_model = Circuit()
     # Input to the model
     shape = [3, 2, 3]
-    w = 0.1*torch.rand(1, *shape, requires_grad=True)
-    x = 0.1*torch.rand(1, *shape, requires_grad=True)
-    y = 0.1*torch.rand(1, *shape, requires_grad=True)
+    w = 2 * torch.rand(1, *shape, requires_grad=True) - 1
+    x = 2 * torch.rand(1, *shape, requires_grad=True) - 1
+    y = 2 * torch.rand(1, *shape, requires_grad=True) - 1
     torch_out = torch_model(w, x, y)
     # Export the model
     torch.onnx.export(torch_model,               # model being run

examples/onnx/rounding_ops/input.json

@@ -1 +1,148 @@
-{"input_shapes": [[3, 2, 3], [3, 2, 3], [3, 2, 3], [3, 2, 3]], "input_data": [[0.0025284828152507544, 0.04976580664515495, 0.025840921327471733, 0.0829394981265068, 0.09595223516225815, 0.08764562010765076, 0.06308566778898239, 0.062386948615312576, 0.08090643584728241, 0.09267748892307281, 0.07428313046693802, 0.08987367898225784, 0.005716216750442982, 0.0666426345705986, 0.012837404385209084, 0.05769496038556099, 0.05761152133345604, 0.08006472885608673], [0.007834953255951405, 0.011380612850189209, 0.08560049533843994, 0.022283583879470825, 0.07879520952701569, 0.04422441124916077, 0.030812596902251244, 0.006081616971641779, 0.011045408435165882, 0.08776585012674332, 0.044985152781009674, 0.015603715553879738, 0.07923348993062973, 0.04872611165046692, 0.0036642670165747404, 0.05142095685005188, 0.0963878259062767, 0.03225792199373245], [0.09952805936336517, 0.002214533044025302, 0.011696457862854004, 0.022422820329666138, 0.04151459410786629, 0.027647346258163452, 0.011919880285859108, 0.006539052817970514, 0.06569185107946396, 0.034328874200582504, 0.0032284557819366455, 0.004105025436729193, 0.022395813837647438, 0.07135921716690063, 0.07882415503263474, 0.09764843434095383, 0.05335796996951103, 0.0525360181927681]], "output_data": [[0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], [0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], [1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0]]}
+{
+    "input_shapes": [
+        [
+            3,
+            2,
+            3
+        ],
+        [
+            3,
+            2,
+            3
+        ],
+        [
+            3,
+            2,
+            3
+        ],
+        [
+            3,
+            2,
+            3
+        ]
+    ],
+    "input_data": [
+        [
+            0.5,
+            1.5,
+            -0.04514765739440918,
+            0.5936200618743896,
+            0.9271858930587769,
+            0.6688600778579712,
+            -0.20331168174743652,
+            -0.7016235589981079,
+            0.025863051414489746,
+            -0.19426143169403076,
+            0.9827852249145508,
+            0.4897397756576538,
+            -1.5,
+            -0.5,
+            0.9278832674026489,
+            0.5943725109100342,
+            -0.573331356048584,
+            0.3675816059112549
+        ],
+        [
+            0.7803324460983276,
+            -0.9616303443908691,
+            0.6070173978805542,
+            -0.028337717056274414,
+            -0.5080242156982422,
+            -0.9280107021331787,
+            0.6150380373001099,
+            0.3865993022918701,
+            -0.43668973445892334,
+            0.17152702808380127,
+            0.5144252777099609,
+            -0.28881049156188965,
+            0.8932310342788696,
+            0.059034109115600586,
+            0.6865451335906982,
+            0.009820222854614258,
+            0.23011493682861328,
+            -0.9492779970169067
+        ],
+        [
+            -0.21352827548980713,
+            -0.16015326976776123,
+            -0.38964390754699707,
+            0.13464701175689697,
+            -0.8814496994018555,
+            0.5037975311279297,
+            -0.804405927658081,
+            0.9858957529067993,
+            0.19567716121673584,
+            0.9777265787124634,
+            0.6151977777481079,
+            0.568595290184021,
+            0.10584986209869385,
+            -0.8975653648376465,
+            0.6235959529876709,
+            -0.547879695892334,
+            0.9289869070053101,
+            0.7567293643951416
+        ]
+    ],
+    "output_data": [
+        [
+            1.0,
+            0.0,
+            -0.0,
+            1.0,
+            1.0,
+            1.0,
+            -0.0,
+            -1.0,
+            0.0,
+            -0.0,
+            1.0,
+            0.0,
+            0.0,
+            1.0,
+            1.0,
+            1.0,
+            -1.0,
+            0.0
+        ],
+        [
+            0.0,
+            -1.0,
+            0.0,
+            -1.0,
+            -1.0,
+            -1.0,
+            0.0,
+            0.0,
+            -1.0,
+            0.0,
+            0.0,
+            -1.0,
+            0.0,
+            0.0,
+            0.0,
+            0.0,
+            0.0,
+            -1.0
+        ],
+        [
+            -0.0,
+            -0.0,
+            -0.0,
+            1.0,
+            -0.0,
+            1.0,
+            -0.0,
+            1.0,
+            1.0,
+            1.0,
+            1.0,
+            1.0,
+            1.0,
+            -0.0,
+            1.0,
+            -0.0,
+            1.0,
+            1.0
+        ]
+    ]
+}

examples/onnx/rounding_ops/network.onnx

@@ -1,10 +1,11 @@
-pytorch2.0.1:â
+pytorch2.2.2:ă
 
 
woutput_w/Round"Round
 
 
xoutput_x/Floor"Floor
 
-
youtput_y/Ceil"Ceil	torch_jitZ%
+
youtput_y/Ceil"Ceil
+main_graphZ%
 
w 
 

 

examples/onnx/rsqrt/gen.py

@@ -0,0 +1,42 @@
+from torch import nn
+import torch
+import json
+import numpy as np
+
+
+class MyModel(nn.Module):
+    def __init__(self):
+        super(MyModel, self).__init__()
+
+    def forward(self, x):         
+        # reciprocal sqrt 
+        m = 1 / torch.sqrt(x)
+        return m 
+
+
+circuit = MyModel()
+
+x = torch.empty(1, 8).uniform_(0, 1)
+
+out = circuit(x)
+
+print(out)
+
+torch.onnx.export(circuit, x, "network.onnx",
+                  export_params=True,        # store the trained parameter weights inside the model file
+                  opset_version=17,          # the ONNX version to export the model to
+                  do_constant_folding=True,  # whether to execute constant folding for optimization
+                  input_names=['input'],   # the model's input names
+                  output_names=['output'],  # the model's output names
+                  dynamic_axes={'input': {0: 'batch_size'},  # variable length axes
+                                'output': {0: 'batch_size'}})
+
+
+d1 = ((x).detach().numpy()).reshape([-1]).tolist()
+
+data = dict(
+    input_data=[d1],
+)
+
+# Serialize data into file:
+json.dump(data, open("input.json", 'w'))

examples/onnx/rsqrt/input.json

@@ -0,0 +1 @@
+{"input_data": [[0.8590779900550842, 0.4029041528701782, 0.6507361531257629, 0.9782488942146301, 0.37392884492874146, 0.6867020726203918, 0.11407750844955444, 0.362740159034729]]}

examples/onnx/rsqrt/network.onnx

@@ -0,0 +1,17 @@
+pytorch2.2.2:Ź

+$
+input/Sqrt_output_0/Sqrt"Sqrt
+1
+/Sqrt_output_0output/Reciprocal"
+Reciprocal
+main_graphZ!
+input
+

+
+batch_size
+b"
+output
+

+
+batch_size
+B

examples/onnx/scatter_nd/gen.py

@@ -1,75 +1,52 @@
-import torch
-import torch.nn as nn
-import sys
+from torch import nn
 import json
-
-sys.path.append("..")
-
-class Model(nn.Module):
-    """
-    Just one Linear layer
-    """
-    def __init__(self, configs):
-        super(Model, self).__init__()
-        self.seq_len = configs.seq_len
-        self.pred_len = configs.pred_len
-
-        # Use this line if you want to visualize the weights
-        # self.Linear.weight = nn.Parameter((1/self.seq_len)*torch.ones([self.pred_len,self.seq_len]))
-        self.channels = configs.enc_in
-        self.individual = configs.individual
-        if self.individual:
-            self.Linear = nn.ModuleList()
-            for i in range(self.channels):
-                self.Linear.append(nn.Linear(self.seq_len,self.pred_len))
-        else:
-            self.Linear = nn.Linear(self.seq_len, self.pred_len)
-
-    def forward(self, x):
-        # x: [Batch, Input length, Channel]
-        if self.individual:
-            output = torch.zeros([x.size(0),self.pred_len,x.size(2)],dtype=x.dtype).to(x.device)
-            for i in range(self.channels):
-                output[:,:,i] = self.Linear[i](x[:,:,i])
-            x = output
-        else:
-            x = self.Linear(x.permute(0,2,1)).permute(0,2,1)
-        return x # [Batch, Output length, Channel]
-
-class Configs:
-    def __init__(self, seq_len, pred_len, enc_in=321, individual=True):
-      self.seq_len = seq_len
-      self.pred_len = pred_len
-      self.enc_in = enc_in
-      self.individual = individual
-
-model = 'Linear'
-seq_len = 10
-pred_len = 4
-enc_in = 3
-
-configs = Configs(seq_len, pred_len, enc_in, True)
-circuit = Model(configs)
-
-x = torch.randn(1, seq_len, pred_len)
+import numpy as np
+import tf2onnx
 
 
-torch.onnx.export(circuit, x, "network.onnx",
-                  export_params=True,        # store the trained parameter weights inside the model file
-                  opset_version=15,          # the ONNX version to export the model to
-                  do_constant_folding=True,  # whether to execute constant folding for optimization
-                  # the model's input names
-                  input_names=['input'],
-                  output_names=['output'],  # the model's output names
-                  dynamic_axes={'input': {0: 'batch_size'},  # variable length axes
-                                'output': {0: 'batch_size'}})
+import tensorflow as tf
+from tensorflow.keras.layers import *
+from tensorflow.keras.models import Model
 
 
-d1 = ((x).detach().numpy()).reshape([-1]).tolist()
+# gather_nd in tf then export to onnx
+x = in1 = Input((4, 1), dtype=tf.int32)
+w = in2 = Input((4, ), dtype=tf.int32)
+
+class MyLayer(Layer):
+    def call(self, x, w):
+        shape = tf.constant([8])
+        return tf.scatter_nd(x, w, shape)
+
+x = MyLayer()(x, w)
+
+
+
+tm = Model((in1, in2), x)
+tm.summary()
+tm.compile(optimizer='adam', loss='mse')
+
+shape = [1, 4, 1]
+index_shape = [1, 4]
+# After training, export to onnx (network.onnx) and create a data file (input.json)
+x = np.random.randint(0, 4, shape)
+# w = random int tensor
+w = np.random.randint(0, 4, index_shape)
+
+spec = tf.TensorSpec(shape, tf.int32, name='input_0')
+index_spec = tf.TensorSpec(index_shape, tf.int32, name='input_1')
+
+model_path = "network.onnx"
+
+tf2onnx.convert.from_keras(tm, input_signature=[spec, index_spec], inputs_as_nchw=['input_0', 'input_1'], opset=12, output_path=model_path)
+
+
+d = x.reshape([-1]).tolist()
+d1 = w.reshape([-1]).tolist()
 
 
 data = dict(
-    input_data=[d1],
+    input_data=[d, d1],
 )
 
 # Serialize data into file:

examples/onnx/scatter_nd/input.json

@@ -1 +1,16 @@
-{"input_data": [[0.1874287724494934, 1.0498261451721191, 0.22384068369865417, 1.048445224761963, -0.5670360326766968, -0.38653188943862915, 0.12878702580928802, -2.3675858974456787, 0.5800458192825317, -0.43653929233551025, -0.2511898875236511, 0.3324051797389984, 0.27960312366485596, 0.4763695001602173, 0.3796705901622772, 1.1334782838821411, -0.87981778383255, -1.2451434135437012, 0.7672272324562073, -0.24404007196426392, -0.6875824928283691, 0.3619358539581299, -0.10131897777318954, 0.7169521450996399, 1.6585893630981445, -0.5451845526695251, 0.429487019777298, 0.7426952123641968, -0.2543637454509735, 0.06546942889690399, 0.7939824461936951, 0.1579471379518509, -0.043604474514722824, -0.8621711730957031, -0.5344759821891785, -0.05880478024482727, -0.17351101338863373, 0.5095029473304749, -0.7864817976951599, -0.449171245098114]]}
+{
+    "input_data": [
+        [
+            0,
+            1,
+            2,
+            3
+        ],
+        [
+            1,
+            0,
+            2,
+            1
+        ]
+    ]
+}

examples/onnx/smallworm/.gitattributes

@@ -0,0 +1 @@
+network.onnx filter=lfs diff=lfs merge=lfs -text

examples/onnx/smallworm/Readme.md

@@ -0,0 +1,47 @@
+## The worm 
+
+This is an onnx file for a [WormVAE](https://github.com/TuragaLab/wormvae?tab=readme-ov-file) model, which is a VAE / latent-space representation of the C. elegans connectome.
+
+The model "is a large-scale latent variable model with a very high-dimensional latent space
+consisting of voltage dynamics of 300 neurons over 5 minutes of time at the simulation frequency
+of 160 Hz. The generative model for these latent variables is described by stochastic differential
+equations modeling the nonlinear dynamics of the network activity." (see [here](https://openreview.net/pdf?id=CJzi3dRlJE-)). 
+
+In effect this is a generative model for a worm's voltage dynamics, which can be used to generate new worm-like voltage dynamics given previous connectome state.
+
+Using ezkl you can create a zk circuit equivalent to the wormvae model, allowing you to "prove" execution of the worm model. If you're feeling particularly adventurous, you can also use the zk circuit to generate new worm-state that can be verified on chain. 
+
+To do so you'll first want to fetch the files using git-lfs (as the onnx file is too large to be stored in git). 
+
+```bash
+git lfs fetch --all
+```
+
+You'll then want to use the usual ezkl loop to generate the zk circuit. We recommend using fixed visibility for the model parameters, as the model is quite large and this will prune the circuit significantly. 
+
+```bash
+ezkl gen-settings --param-visibility=fixed
+cp input.json calibration.json
+ezkl calibrate-settings
+ezkl compile-circuit
+ezkl gen-witness
+ezkl prove
+```
+
+You might also need to aggregate the proof to get it to fit on chain.
+
+```bash
+ezkl aggregate
+```
+
+You can then create a smart contract that verifies this aggregate proof
+
+```bash
+ezkl create-evm-verifier-aggr
+```
+
+This can then be deployed on the chain of your choice.
+
+
+> Note: the model is large and thus we recommend a machine with at least 512GB of RAM to run the above commands. If you're ever compute constrained you can always use the lilith service to generate the zk circuit. Message us on discord or telegram for more details :)
+

examples/onnx/smallworm/network.onnx

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2f88c5901d3768ec21e3cf2f2840d255e84fa13c364df86b24d960cca3333769
+size 82095882

examples/onnx/smallworm/settings.json

@@ -0,0 +1 @@
+{"run_args":{"tolerance":{"val":0.0,"scale":1.0},"input_scale":0,"param_scale":6,"scale_rebase_multiplier":1,"lookup_range":[-32768,32768],"logrows":17,"variables":[["batch_size",1]],"input_visibility":"Private","output_visibility":"Public","param_visibility":"Fixed"},"num_constraints":367422820,"total_const_size":365577160,"model_instance_shapes":[[1,300,1200]],"model_output_scales":[6],"model_input_scales":[0,0,0],"module_sizes":{"kzg":[],"poseidon":[0,[0]],"elgamal":[0,[0]]},"required_lookups":[{"Div":{"denom":64.0}},"ReLU",{"Ln":{"scale":64.0}},{"Exp":{"scale":64.0}}],"check_mode":"UNSAFE","version":"0.0.0","num_blinding_factors":null}

ezkl.pyi

@@ -0,0 +1,851 @@
+# This file is automatically generated by pyo3_stub_gen
+# ruff: noqa: E501, F401
+
+import os
+import pathlib
+import typing
+from enum import Enum, auto
+
+class PyG1:
+    r"""
+    pyclass containing the struct used for G1, this is mostly a helper class
+    """
+    ...
+
+class PyG1Affine:
+    r"""
+    pyclass containing the struct used for G1
+    """
+    ...
+
+class PyRunArgs:
+    r"""
+    Python class containing the struct used for run_args
+    
+    Returns
+    -------
+    PyRunArgs
+    """
+    ...
+
+class PyCommitments(Enum):
+    r"""
+    pyclass representing an enum, denoting the type of commitment
+    """
+    KZG = auto()
+    IPA = auto()
+
+class PyInputType(Enum):
+    Bool = auto()
+    F16 = auto()
+    F32 = auto()
+    F64 = auto()
+    Int = auto()
+    TDim = auto()
+
+class PyTestDataSource(Enum):
+    r"""
+    pyclass representing an enum
+    """
+    File = auto()
+    OnChain = auto()
+
+def aggregate(aggregation_snarks:typing.Sequence[str | os.PathLike | pathlib.Path],proof_path:str | os.PathLike | pathlib.Path,vk_path:str | os.PathLike | pathlib.Path,transcript:str,logrows:int,check_mode:str,split_proofs:bool,srs_path:typing.Optional[str | os.PathLike | pathlib.Path],commitment:PyCommitments) -> bool:
+    r"""
+    Creates an aggregated proof
+    
+    Arguments
+    ---------
+    aggregation_snarks: list[str]
+        List of paths to the various proofs
+    
+    proof_path: str
+        Path to output the aggregated proof
+    
+    vk_path: str
+        Path to the VK file
+    
+    transcript:
+        Proof transcript type to be used. `evm` used by default. `poseidon` is also supported
+    
+    logrows:
+        Logrows used for aggregation circuit
+    
+    check_mode: str
+        Run sanity checks during calculations. Accepts `safe` or `unsafe`
+    
+    split-proofs: bool
+         Whether the accumulated proofs are segments of a larger circuit
+    
+    srs_path: str
+        Path to the SRS used
+    
+    commitment: str
+        Accepts "kzg" or "ipa"
+    
+    Returns
+    -------
+    bool
+    """
+    ...
+
+def buffer_to_felts(buffer:typing.Sequence[int]) -> list[str]:
+    r"""
+    Converts a buffer to vector of field elements
+    
+    Arguments
+    -------
+    buffer: list[int]
+        List of integers representing a buffer
+    
+    Returns
+    -------
+    list[str]
+        List of field elements represented as strings
+    """
+    ...
+
+def calibrate_settings(data:str | os.PathLike | pathlib.Path,model:str | os.PathLike | pathlib.Path,settings:str | os.PathLike | pathlib.Path,target:str,lookup_safety_margin:float,scales:typing.Optional[typing.Sequence[int]],scale_rebase_multiplier:typing.Sequence[int],max_logrows:typing.Optional[int]) -> typing.Any:
+    r"""
+    Calibrates the circuit settings
+    
+    Arguments
+    ---------
+    data: str
+        Path to the calibration data
+    
+    model: str
+        Path to the onnx file
+    
+    settings: str
+        Path to the settings file
+    
+    lookup_safety_margin: int
+         the lookup safety margin to use for calibration. if the max lookup is 2^k, then the max lookup will be 2^k * lookup_safety_margin. larger = safer but slower
+    
+    scales: list[int]
+        Optional scales to specifically try for calibration
+    
+    scale_rebase_multiplier: list[int]
+        Optional scale rebase multipliers to specifically try for calibration. This is the multiplier at which we divide to return to the input scale.
+    
+    max_logrows: int
+        Optional max logrows to use for calibration
+    
+    
+    Returns
+    -------
+    bool
+    """
+    ...
+
+def compile_circuit(model:str | os.PathLike | pathlib.Path,compiled_circuit:str | os.PathLike | pathlib.Path,settings_path:str | os.PathLike | pathlib.Path) -> bool:
+    r"""
+    Compiles the circuit for use in other steps
+    
+    Arguments
+    ---------
+    model: str
+        Path to the onnx model file
+    
+    compiled_circuit: str
+        Path to output the compiled circuit
+    
+    settings_path: str
+        Path to the settings files
+    
+    Returns
+    -------
+    bool
+    """
+    ...
+
+def create_evm_data_attestation(input_data:str | os.PathLike | pathlib.Path,settings_path:str | os.PathLike | pathlib.Path,sol_code_path:str | os.PathLike | pathlib.Path,abi_path:str | os.PathLike | pathlib.Path,witness_path:typing.Optional[str | os.PathLike | pathlib.Path]) -> typing.Any:
+    r"""
+    Creates an EVM compatible data attestation verifier, you will need solc installed in your environment to run this
+    
+    Arguments
+    ---------
+    input_data: str
+        The path to the .json data file, which should contain the necessary calldata and account addresses needed to read from all the on-chain view functions that return the data that the network ingests as inputs
+    
+    settings_path: str
+        The path to the settings file
+    
+    sol_code_path: str
+        The path to the create the solidity verifier
+    
+    abi_path: str
+        The path to create the ABI for the solidity verifier
+    
+    Returns
+    -------
+    bool
+    """
+    ...
+
+def create_evm_verifier(vk_path:str | os.PathLike | pathlib.Path,settings_path:str | os.PathLike | pathlib.Path,sol_code_path:str | os.PathLike | pathlib.Path,abi_path:str | os.PathLike | pathlib.Path,srs_path:typing.Optional[str | os.PathLike | pathlib.Path],reusable:bool) -> typing.Any:
+    r"""
+    Creates an EVM compatible verifier, you will need solc installed in your environment to run this
+    
+    Arguments
+    ---------
+    vk_path: str
+        The path to the verification key file
+    
+    settings_path: str
+        The path to the settings file
+    
+    sol_code_path: str
+        The path to the create the solidity verifier
+    
+    abi_path: str
+        The path to create the ABI for the solidity verifier
+    
+    srs_path: str
+        The path to the SRS file
+    
+    reusable: bool
+        Whether the verifier should be rendered as a reusable contract. If so, then you will need to deploy the VK artifact separately which you can generate using the create_evm_vka command
+    
+    Returns
+    -------
+    bool
+    """
+    ...
+
+def create_evm_verifier_aggr(aggregation_settings:typing.Sequence[str | os.PathLike | pathlib.Path],vk_path:str | os.PathLike | pathlib.Path,sol_code_path:str | os.PathLike | pathlib.Path,abi_path:str | os.PathLike | pathlib.Path,logrows:int,srs_path:typing.Optional[str | os.PathLike | pathlib.Path],reusable:bool) -> typing.Any:
+    r"""
+    Creates an evm compatible aggregate verifier, you will need solc installed in your environment to run this
+    
+    Arguments
+    ---------
+    aggregation_settings: str
+        path to the settings file
+    
+    vk_path: str
+        The path to load the desired verification key file
+    
+    sol_code_path: str
+        The path to the Solidity code
+    
+    abi_path: str
+        The path to output the Solidity verifier ABI
+    
+    logrows: int
+        Number of logrows used during aggregated setup
+    
+    srs_path: str
+        The path to the SRS file
+    
+    reusable: bool
+        Whether the verifier should be rendered as a reusable contract. If so, then you will need to deploy the VK artifact separately which you can generate using the create_evm_vka command
+    
+    Returns
+    -------
+    bool
+    """
+    ...
+
+def create_evm_vka(vk_path:str | os.PathLike | pathlib.Path,settings_path:str | os.PathLike | pathlib.Path,sol_code_path:str | os.PathLike | pathlib.Path,abi_path:str | os.PathLike | pathlib.Path,srs_path:typing.Optional[str | os.PathLike | pathlib.Path]) -> typing.Any:
+    r"""
+    Creates an Evm VK artifact. This command generated a VK with circuit specific meta data encoding in memory for use by the reusable H2 verifier.
+    This is useful for deploying verifier that were otherwise too big to fit on chain and required aggregation.
+    
+    Arguments
+    ---------
+    vk_path: str
+        The path to the verification key file
+    
+    settings_path: str
+        The path to the settings file
+    
+    sol_code_path: str
+        The path to the create the solidity verifying key.
+    
+    abi_path: str
+        The path to create the ABI for the solidity verifier
+    
+    srs_path: str
+        The path to the SRS file
+    
+    Returns
+    -------
+    bool
+    """
+    ...
+
+def deploy_da_evm(addr_path:str | os.PathLike | pathlib.Path,input_data:str | os.PathLike | pathlib.Path,settings_path:str | os.PathLike | pathlib.Path,sol_code_path:str | os.PathLike | pathlib.Path,rpc_url:typing.Optional[str],optimizer_runs:int,private_key:typing.Optional[str]) -> typing.Any:
+    r"""
+    deploys the solidity da verifier
+    """
+    ...
+
+def deploy_evm(addr_path:str | os.PathLike | pathlib.Path,sol_code_path:str | os.PathLike | pathlib.Path,rpc_url:typing.Optional[str],contract_type:str,optimizer_runs:int,private_key:typing.Optional[str]) -> typing.Any:
+    r"""
+    deploys the solidity verifier
+    """
+    ...
+
+def encode_evm_calldata(proof:str | os.PathLike | pathlib.Path,calldata:str | os.PathLike | pathlib.Path,addr_vk:typing.Optional[str]) -> list[int]:
+    r"""
+    Creates encoded evm calldata from a proof file
+    
+    Arguments
+    ---------
+    proof: str
+        Path to the proof file
+    
+    calldata: str
+       Path to the calldata file to save
+    
+    addr_vk: str
+       The address of the verification key contract (if the verifier key is to be rendered as a separate contract)
+    
+    Returns
+    -------
+    vec[u8]
+       The encoded calldata
+    """
+    ...
+
+def felt_to_big_endian(felt:str) -> str:
+    r"""
+    Converts a field element hex string to big endian
+    
+    Arguments
+    -------
+    felt: str
+        The field element represented as a string
+    
+    
+    Returns
+    -------
+    str
+        field element represented as a string
+    """
+    ...
+
+def felt_to_float(felt:str,scale:int) -> float:
+    r"""
+    Converts a field element hex string to a floating point number
+    
+    Arguments
+    -------
+    felt: str
+       The field element represented as a string
+    
+    scale: float
+        The scaling factor used to convert the field element into a floating point representation
+    
+    Returns
+    -------
+    float
+    """
+    ...
+
+def felt_to_int(felt:str) -> int:
+    r"""
+    Converts a field element hex string to an integer
+    
+    Arguments
+    -------
+    felt: str
+        The field element represented as a string
+    
+    Returns
+    -------
+    int
+    """
+    ...
+
+def float_to_felt(input:float,scale:int,input_type:PyInputType) -> str:
+    r"""
+    Converts a floating point element to a field element hex string
+    
+    Arguments
+    -------
+    input: float
+       The field element represented as a string
+    
+    scale: float
+        The scaling factor used to quantize the float into a field element
+    
+    input_type: PyInputType
+        The type of the input
+    
+    Returns
+    -------
+    str
+        The field element represented as a string
+    """
+    ...
+
+def gen_settings(model:str | os.PathLike | pathlib.Path,output:str | os.PathLike | pathlib.Path,py_run_args:typing.Optional[PyRunArgs]) -> bool:
+    r"""
+    Generates the circuit settings
+    
+    Arguments
+    ---------
+    model: str
+        Path to the onnx file
+    
+    output: str
+        Path to create the settings file
+    
+    py_run_args: PyRunArgs
+        PyRunArgs object to initialize the settings
+    
+    Returns
+    -------
+    bool
+    """
+    ...
+
+def gen_srs(srs_path:str | os.PathLike | pathlib.Path,logrows:int) -> None:
+    r"""
+    Generates the Structured Reference String (SRS), use this only for testing purposes
+    
+    Arguments
+    ---------
+    srs_path: str
+        Path to the create the SRS file
+    
+    logrows: int
+        The number of logrows for the SRS file
+    """
+    ...
+
+def gen_vk_from_pk_aggr(path_to_pk:str | os.PathLike | pathlib.Path,vk_output_path:str | os.PathLike | pathlib.Path) -> bool:
+    r"""
+    Generates a vk from a pk for an aggregate circuit and saves it to a file
+    
+    Arguments
+    -------
+    path_to_pk: str
+        Path to the proving key
+    
+    vk_output_path: str
+        Path to create the vk file
+    
+    Returns
+    -------
+    bool
+    """
+    ...
+
+def gen_vk_from_pk_single(path_to_pk:str | os.PathLike | pathlib.Path,circuit_settings_path:str | os.PathLike | pathlib.Path,vk_output_path:str | os.PathLike | pathlib.Path) -> bool:
+    r"""
+    Generates a vk from a pk for a model circuit and saves it to a file
+    
+    Arguments
+    -------
+    path_to_pk: str
+        Path to the proving key
+    
+    circuit_settings_path: str
+        Path to the witness file
+    
+    vk_output_path: str
+        Path to create the vk file
+    
+    Returns
+    -------
+    bool
+    """
+    ...
+
+def gen_witness(data:str | os.PathLike | pathlib.Path,model:str | os.PathLike | pathlib.Path,output:typing.Optional[str | os.PathLike | pathlib.Path],vk_path:typing.Optional[str | os.PathLike | pathlib.Path],srs_path:typing.Optional[str | os.PathLike | pathlib.Path]) -> typing.Any:
+    r"""
+    Runs the forward pass operation to generate a witness
+    
+    Arguments
+    ---------
+    data: str
+        Path to the data file
+    
+    model: str
+        Path to the compiled model file
+    
+    output: str
+        Path to create the witness file
+    
+    vk_path: str
+        Path to the verification key
+    
+    srs_path: str
+        Path to the SRS file
+    
+    Returns
+    -------
+    dict
+        Python object containing the witness values
+    """
+    ...
+
+def get_srs(settings_path:typing.Optional[str | os.PathLike | pathlib.Path],logrows:typing.Optional[int],srs_path:typing.Optional[str | os.PathLike | pathlib.Path],commitment:typing.Optional[PyCommitments]) -> typing.Any:
+    r"""
+    Gets a public srs
+    
+    Arguments
+    ---------
+    settings_path: str
+        Path to the settings file
+    
+    logrows: int
+        The number of logrows for the SRS file
+    
+    srs_path: str
+        Path to the create the SRS file
+    
+    commitment: str
+        Specify the commitment used ("kzg", "ipa")
+    
+    Returns
+    -------
+    bool
+    """
+    ...
+
+def ipa_commit(message:typing.Sequence[str],vk_path:str | os.PathLike | pathlib.Path,settings_path:str | os.PathLike | pathlib.Path,srs_path:typing.Optional[str | os.PathLike | pathlib.Path]) -> list[PyG1Affine]:
+    r"""
+    Generate an ipa commitment.
+    
+    Arguments
+    -------
+    message: list[str]
+        List of field elements represnted as strings
+    
+    vk_path: str
+        Path to the verification key
+    
+    settings_path: str
+        Path to the settings file
+    
+    srs_path: str
+        Path to the Structure Reference String (SRS) file
+    
+    Returns
+    -------
+    list[PyG1Affine]
+    """
+    ...
+
+def kzg_commit(message:typing.Sequence[str],vk_path:str | os.PathLike | pathlib.Path,settings_path:str | os.PathLike | pathlib.Path,srs_path:typing.Optional[str | os.PathLike | pathlib.Path]) -> list[PyG1Affine]:
+    r"""
+    Generate a kzg commitment.
+    
+    Arguments
+    -------
+    message: list[str]
+        List of field elements represnted as strings
+    
+    vk_path: str
+        Path to the verification key
+    
+    settings_path: str
+        Path to the settings file
+    
+    srs_path: str
+        Path to the Structure Reference String (SRS) file
+    
+    Returns
+    -------
+    list[PyG1Affine]
+    """
+    ...
+
+def mock(witness:str | os.PathLike | pathlib.Path,model:str | os.PathLike | pathlib.Path) -> bool:
+    r"""
+    Mocks the prover
+    
+    Arguments
+    ---------
+    witness: str
+        Path to the witness file
+    
+    model: str
+        Path to the compiled model file
+    
+    Returns
+    -------
+    bool
+    """
+    ...
+
+def mock_aggregate(aggregation_snarks:typing.Sequence[str | os.PathLike | pathlib.Path],logrows:int,split_proofs:bool) -> bool:
+    r"""
+    Mocks the aggregate prover
+    
+    Arguments
+    ---------
+    aggregation_snarks: list[str]
+        List of paths to the relevant proof files
+    
+    logrows: int
+        Number of logrows to use for the aggregation circuit
+    
+    split_proofs: bool
+        Indicates whether the accumulated are segments of a larger proof
+    
+    Returns
+    -------
+    bool
+    """
+    ...
+
+def poseidon_hash(message:typing.Sequence[str]) -> list[str]:
+    r"""
+    Generate a poseidon hash.
+    
+    Arguments
+    -------
+    message: list[str]
+        List of field elements represented as strings
+    
+    Returns
+    -------
+    list[str]
+        List of field elements represented as strings
+    """
+    ...
+
+def prove(witness:str | os.PathLike | pathlib.Path,model:str | os.PathLike | pathlib.Path,pk_path:str | os.PathLike | pathlib.Path,proof_path:typing.Optional[str | os.PathLike | pathlib.Path],proof_type:str,srs_path:typing.Optional[str | os.PathLike | pathlib.Path]) -> typing.Any:
+    r"""
+    Runs the prover on a set of inputs
+    
+    Arguments
+    ---------
+    witness: str
+        Path to the witness file
+    
+    model: str
+        Path to the compiled model file
+    
+    pk_path: str
+        Path to the proving key file
+    
+    proof_path: str
+        Path to create the proof file
+    
+    proof_type: str
+        Accepts `single`, `for-aggr`
+    
+    srs_path: str
+        Path to the SRS file
+    
+    Returns
+    -------
+    bool
+    """
+    ...
+
+def setup(model:str | os.PathLike | pathlib.Path,vk_path:str | os.PathLike | pathlib.Path,pk_path:str | os.PathLike | pathlib.Path,srs_path:typing.Optional[str | os.PathLike | pathlib.Path],witness_path:typing.Optional[str | os.PathLike | pathlib.Path],disable_selector_compression:bool) -> bool:
+    r"""
+    Runs the setup process
+    
+    Arguments
+    ---------
+    model: str
+        Path to the compiled model file
+    
+    vk_path: str
+        Path to create the verification key file
+    
+    pk_path: str
+        Path to create the proving key file
+    
+    srs_path: str
+        Path to the SRS file
+    
+    witness_path: str
+        Path to the witness file
+    
+    disable_selector_compression: bool
+        Whether to compress the selectors or not
+    
+    Returns
+    -------
+    bool
+    """
+    ...
+
+def setup_aggregate(sample_snarks:typing.Sequence[str | os.PathLike | pathlib.Path],vk_path:str | os.PathLike | pathlib.Path,pk_path:str | os.PathLike | pathlib.Path,logrows:int,split_proofs:bool,srs_path:typing.Optional[str | os.PathLike | pathlib.Path],disable_selector_compression:bool,commitment:PyCommitments) -> bool:
+    r"""
+     Runs the setup process for an aggregate setup
+    
+    Arguments
+    ---------
+    sample_snarks: list[str]
+        List of paths to the various proofs
+    
+    vk_path: str
+        Path to create the aggregated VK
+    
+    pk_path: str
+        Path to create the aggregated PK
+    
+    logrows: int
+        Number of logrows to use
+    
+    split_proofs: bool
+        Whether the accumulated are segments of a larger proof
+    
+    srs_path: str
+        Path to the SRS file
+    
+    disable_selector_compression: bool
+        Whether to compress selectors
+    
+    commitment: str
+        Accepts `kzg`, `ipa`
+    
+    Returns
+    -------
+    bool
+    """
+    ...
+
+def setup_test_evm_witness(data_path:str | os.PathLike | pathlib.Path,compiled_circuit_path:str | os.PathLike | pathlib.Path,test_data:str | os.PathLike | pathlib.Path,input_source:PyTestDataSource,output_source:PyTestDataSource,rpc_url:typing.Optional[str]) -> typing.Any:
+    r"""
+    Setup test evm witness
+    
+    Arguments
+    ---------
+    data_path: str
+        The path to the .json data file, which should include both the network input (possibly private) and the network output (public input to the proof)
+    
+    compiled_circuit_path: str
+        The path to the compiled model file (generated using the compile-circuit command)
+    
+    test_data: str
+        For testing purposes only. The optional path to the .json data file that will be generated that contains the OnChain data storage information derived from the file information in the data .json file. Should include both the network input (possibly private) and the network output (public input to the proof)
+    
+    input_sources: str
+        Where the input data comes from
+    
+    output_source: str
+        Where the output data comes from
+    
+    rpc_url: str
+        RPC URL for an EVM compatible node, if None, uses Anvil as a local RPC node
+    
+    Returns
+    -------
+    bool
+    """
+    ...
+
+def swap_proof_commitments(proof_path:str | os.PathLike | pathlib.Path,witness_path:str | os.PathLike | pathlib.Path) -> None:
+    r"""
+    Swap the commitments in a proof
+    
+    Arguments
+    -------
+    proof_path: str
+        Path to the proof file
+    
+    witness_path: str
+        Path to the witness file
+    """
+    ...
+
+def table(model:str | os.PathLike | pathlib.Path,py_run_args:typing.Optional[PyRunArgs]) -> str:
+    r"""
+    Displays the table as a string in python
+    
+    Arguments
+    ---------
+    model: str
+        Path to the onnx file
+    
+    Returns
+    ---------
+    str
+        Table of the nodes in the onnx file
+    """
+    ...
+
+def verify(proof_path:str | os.PathLike | pathlib.Path,settings_path:str | os.PathLike | pathlib.Path,vk_path:str | os.PathLike | pathlib.Path,srs_path:typing.Optional[str | os.PathLike | pathlib.Path],reduced_srs:bool) -> bool:
+    r"""
+    Verifies a given proof
+    
+    Arguments
+    ---------
+    proof_path: str
+        Path to create the proof file
+    
+    settings_path: str
+        Path to the settings file
+    
+    vk_path: str
+        Path to the verification key file
+    
+    srs_path: str
+        Path to the SRS file
+    
+    non_reduced_srs: bool
+        Whether to reduce the number of SRS logrows to the number of instances rather than the number of logrows used for proofs (only works if the srs were generated in the same ceremony)
+    
+    Returns
+    -------
+    bool
+    """
+    ...
+
+def verify_aggr(proof_path:str | os.PathLike | pathlib.Path,vk_path:str | os.PathLike | pathlib.Path,logrows:int,commitment:PyCommitments,reduced_srs:bool,srs_path:typing.Optional[str | os.PathLike | pathlib.Path]) -> bool:
+    r"""
+    Verifies and aggregate proof
+    
+    Arguments
+    ---------
+    proof_path: str
+         The path to the proof file
+    
+    vk_path: str
+        The path to the verification key file
+    
+    logrows: int
+        logrows used for aggregation circuit
+    
+    commitment: str
+        Accepts "kzg" or "ipa"
+    
+    reduced_srs: bool
+        Whether to reduce the number of SRS logrows to the number of instances rather than the number of logrows used for proofs (only works if the srs were generated in the same ceremony)
+    
+    srs_path: str
+        The path to the SRS file
+    
+    Returns
+    -------
+    bool
+    """
+    ...
+
+def verify_evm(addr_verifier:str,proof_path:str | os.PathLike | pathlib.Path,rpc_url:typing.Optional[str],addr_da:typing.Optional[str],addr_vk:typing.Optional[str]) -> typing.Any:
+    r"""
+    verifies an evm compatible proof, you will need solc installed in your environment to run this
+    
+    Arguments
+    ---------
+    addr_verifier: str
+        The verifier contract's address as a hex string
+    
+    proof_path: str
+        The path to the proof file (generated using the prove command)
+    
+    rpc_url: str
+        RPC URL for an Ethereum node, if None will use Anvil but WON'T persist state
+    
+    addr_da: str
+        does the verifier use data attestation ?
+    
+    addr_vk: str
+        The addess of the separate VK contract (if the verifier key is rendered as a separate contract)
+    Returns
+    -------
+    bool
+    """
+    ...
+

in-browser-evm-verifier/src/index.ts

@@ -9,7 +9,6 @@ import { EVM } from '@ethereumjs/evm'
 import { buildTransaction, encodeDeployment } from './utils/tx-builder'
 import { getAccountNonce, insertAccount } from './utils/account-utils'
 import { encodeVerifierCalldata } from '../nodejs/ezkl';
-import { error } from 'console'
 
 async function deployContract(
   vm: VM,
@@ -66,7 +65,7 @@ async function verify(
     vkAddress = new Uint8Array(uint8Array.buffer);
 
     // convert uitn8array of length
-    error('vkAddress', vkAddress)
+    console.error('vkAddress', vkAddress)
   }
   const data = encodeVerifierCalldata(proof, vkAddress)
 

pyproject.toml

@@ -12,6 +12,7 @@ asyncio_mode = "auto"
 
 [project]
 name = "ezkl"
+version = "0.0.0"
 requires-python = ">=3.7"
 classifiers = [
     "Programming Language :: Rust",

rust-toolchain

@@ -1,3 +1,3 @@
 [toolchain]
-channel = "nightly-2024-02-06"
+channel = "nightly-2024-07-18"
 components = ["rustfmt", "clippy"]