ci: update version string in docs

---------

Co-authored-by: dante <45801863+alexander-camuto@users.noreply.github.com>
Co-authored-by: Ethan Cemer <tylercemer@gmail.com>
Co-authored-by: github-actions[bot] <github-actions[bot]@users.noreply.github.com>

.github/workflows/large-tests.yml

@@ -11,7 +11,7 @@ jobs:
       - uses: actions/checkout@v4
       - uses: actions-rs/toolchain@v1
         with:
-          toolchain: nightly-2024-01-04
+          toolchain: nightly-2024-02-06
           override: true
           components: rustfmt, clippy
       - name: nanoGPT Mock

.github/workflows/npm.yml

@@ -22,7 +22,7 @@ jobs:
       - uses: actions/checkout@v4
       - uses: actions-rs/toolchain@v1
         with:
-          toolchain: nightly-2024-01-04
+          toolchain: nightly-2024-02-06
           override: true
           components: rustfmt, clippy
       - uses: jetli/wasm-pack-action@v0.4.0
@@ -30,13 +30,13 @@ jobs:
         run: rustup target add wasm32-unknown-unknown
 
       - name: Add rust-src
-        run: rustup component add rust-src --toolchain nightly-2024-01-04-x86_64-unknown-linux-gnu
+        run: rustup component add rust-src --toolchain nightly-2024-02-06-x86_64-unknown-linux-gnu
       - name: Install binaryen
         run: |
-            set -e
-            curl -L https://github.com/WebAssembly/binaryen/releases/download/version_116/binaryen-version_116-x86_64-linux.tar.gz | tar xzf -
-            export PATH=$PATH:$PWD/binaryen-version_116/bin
-            wasm-opt --version
+          set -e
+          curl -L https://github.com/WebAssembly/binaryen/releases/download/version_116/binaryen-version_116-x86_64-linux.tar.gz | tar xzf -
+          export PATH=$PATH:$PWD/binaryen-version_116/bin
+          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"
@@ -92,7 +92,7 @@ jobs:
             const jsonObject = JSONBig.parse(string);
             return jsonObject;
           }
-          
+
           function serialize(data) { // data is an object // return a Uint8ClampedArray
             // Step 1: Stringify the Object with BigInt support
             if (typeof data === "object") {
@@ -100,11 +100,11 @@ jobs:
             }
             // Step 2: Encode the JSON String
             const uint8Array = new TextEncoder().encode(data);
-          
+
             // Step 3: Convert to Uint8ClampedArray
             return new Uint8ClampedArray(uint8Array.buffer);
           }
-          
+
           module.exports = {
             deserialize,
             serialize
@@ -123,7 +123,7 @@ jobs:
             const jsonObject = parse(string);
             return jsonObject;
           }
-          
+
           export function serialize(data) { // data is an object // return a Uint8ClampedArray
             // Step 1: Stringify the Object with BigInt support
             if (typeof data === "object") {
@@ -131,7 +131,7 @@ jobs:
             }
             // Step 2: Encode the JSON String
             const uint8Array = new TextEncoder().encode(data);
-          
+
             // Step 3: Convert to Uint8ClampedArray
             return new Uint8ClampedArray(uint8Array.buffer);
           }

.github/workflows/pypi-gpu.yml

@@ -26,7 +26,7 @@ jobs:
       - uses: actions/checkout@v4
       - uses: actions/setup-python@v4
         with:
-          python-version: 3.7
+          python-version: 3.12
           architecture: x64
 
       - name: Set pyproject.toml version to match github tag

.github/workflows/pypi.yml

@@ -25,7 +25,7 @@ jobs:
       - uses: actions/checkout@v4
       - uses: actions/setup-python@v4
         with:
-          python-version: 3.7
+          python-version: 3.12
           architecture: x64
 
       - name: Set Cargo.toml version to match github tag
@@ -70,7 +70,7 @@ jobs:
       - uses: actions/checkout@v4
       - uses: actions/setup-python@v4
         with:
-          python-version: 3.7
+          python-version: 3.12
           architecture: ${{ matrix.target }}
 
       - name: Set Cargo.toml version to match github tag
@@ -115,7 +115,7 @@ jobs:
       - uses: actions/checkout@v4
       - uses: actions/setup-python@v4
         with:
-          python-version: 3.7
+          python-version: 3.12
           architecture: x64
 
       - name: Set Cargo.toml version to match github tag
@@ -128,6 +128,7 @@ 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: |
@@ -139,6 +140,20 @@ jobs:
           target: ${{ matrix.target }}
           manylinux: auto
           args: --release --out dist --features python-bindings
+          before-script-linux: |
+            # If we're running on rhel centos, install needed packages.
+            if command -v yum &> /dev/null; then
+                yum update -y && yum install -y perl-core openssl openssl-devel pkgconfig libatomic
+
+                # If we're running on i686 we need to symlink libatomic
+                # in order to build openssl with -latomic flag.
+                if [[ ! -d "/usr/lib64" ]]; then
+                    ln -s /usr/lib/libatomic.so.1 /usr/lib/libatomic.so
+                fi
+            else
+                # If we're running on debian-based system.
+                apt update -y && apt-get install -y libssl-dev openssl pkg-config
+            fi
 
       - name: Install built wheel
         if: matrix.target == 'x86_64'
@@ -162,7 +177,7 @@ jobs:
   #   - uses: actions/checkout@v4
   #   - uses: actions/setup-python@v4
   #     with:
-  #       python-version: 3.7
+  #       python-version: 3.12
 
   #   - name: Install cross-compilation tools for aarch64
   #     if: matrix.target == 'aarch64'
@@ -214,7 +229,7 @@ jobs:
       - uses: actions/checkout@v4
       - uses: actions/setup-python@v4
         with:
-          python-version: 3.7
+          python-version: 3.12
           architecture: x64
 
       - name: Set Cargo.toml version to match github tag
@@ -249,7 +264,7 @@ jobs:
             apk add py3-pip
             pip3 install -U pip
             python3 -m venv .venv
-            source .venv/bin/activate            
+            source .venv/bin/activate
             pip3 install ezkl --no-index --find-links /io/dist/ --force-reinstall
             python3 -c "import ezkl"
 
@@ -273,7 +288,7 @@ jobs:
       - uses: actions/checkout@v4
       - uses: actions/setup-python@v4
         with:
-          python-version: 3.7
+          python-version: 3.12
 
       - name: Set Cargo.toml version to match github tag
         shell: bash
@@ -345,3 +360,17 @@ jobs:
         with:
           repository-url: https://test.pypi.org/legacy/
           packages-dir: ./
+
+  doc-publish:
+    name: Trigger ReadTheDocs Build
+    runs-on: ubuntu-latest
+    needs: pypi-publish
+    steps:
+      - uses: actions/checkout@v4
+
+      - 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 }}

.github/workflows/release.yml

@@ -32,7 +32,7 @@ jobs:
           token: ${{ secrets.RELEASE_TOKEN }}
           tag_name: ${{ env.EZKL_VERSION }}
 
-  build-release-gpu: 
+  build-release-gpu:
     name: build-release-gpu
     needs: ["create-release"]
     runs-on: GPU
@@ -45,7 +45,7 @@ jobs:
     steps:
       - uses: actions-rs/toolchain@v1
         with:
-          toolchain: nightly-2024-01-04
+          toolchain: nightly-2024-02-06
           override: true
           components: rustfmt, clippy
       - name: Checkout repo
@@ -60,16 +60,15 @@ jobs:
       - name: Set Cargo.toml version to match github tag
         shell: bash
         run: |
-            mv Cargo.toml Cargo.toml.orig
-            sed "s/0\\.0\\.0/${EZKL_VERSION//v}/" Cargo.toml.orig >Cargo.toml
-            mv Cargo.lock Cargo.lock.orig
-            sed "s/0\\.0\\.0/${EZKL_VERSION//v}/" Cargo.lock.orig >Cargo.lock
-
+          mv Cargo.toml Cargo.toml.orig
+          sed "s/0\\.0\\.0/${EZKL_VERSION//v}/" Cargo.toml.orig >Cargo.toml
+          mv Cargo.lock Cargo.lock.orig
+          sed "s/0\\.0\\.0/${EZKL_VERSION//v}/" Cargo.lock.orig >Cargo.lock
 
       - name: Install dependencies
         shell: bash
         run: |
-            sudo apt-get update
+          sudo apt-get update
 
       - name: Build release binary
         run: cargo build --release -Z sparse-registry --features icicle
@@ -91,7 +90,6 @@ jobs:
           asset_name: ${{ env.ASSET }}
           asset_content_type: application/octet-stream
 
-
   build-release:
     name: build-release
     needs: ["create-release"]

.github/workflows/rust.yml

@@ -26,7 +26,7 @@ jobs:
       - uses: actions/checkout@v4
       - uses: actions-rs/toolchain@v1
         with:
-          toolchain: nightly-2024-01-04
+          toolchain: nightly-2024-02-06
           override: true
           components: rustfmt, clippy
       - name: Build
@@ -38,7 +38,7 @@ jobs:
       - uses: actions/checkout@v4
       - uses: actions-rs/toolchain@v1
         with:
-          toolchain: nightly-2024-01-04
+          toolchain: nightly-2024-02-06
           override: true
           components: rustfmt, clippy
       - name: Docs
@@ -50,7 +50,7 @@ jobs:
       - uses: actions/checkout@v4
       - uses: actions-rs/toolchain@v1
         with:
-          toolchain: nightly-2024-01-04
+          toolchain: nightly-2024-02-06
           override: true
           components: rustfmt, clippy
       - uses: baptiste0928/cargo-install@v1
@@ -73,7 +73,7 @@ jobs:
       - uses: actions/checkout@v4
       - uses: actions-rs/toolchain@v1
         with:
-          toolchain: nightly-2024-01-04
+          toolchain: nightly-2024-02-06
           override: true
           components: rustfmt, clippy
       - uses: baptiste0928/cargo-install@v1
@@ -106,7 +106,7 @@ jobs:
       - uses: actions/checkout@v4
       - uses: actions-rs/toolchain@v1
         with:
-          toolchain: nightly-2024-01-04
+          toolchain: nightly-2024-02-06
           override: true
           components: rustfmt, clippy
       - uses: baptiste0928/cargo-install@v1
@@ -139,7 +139,7 @@ jobs:
       - uses: actions/checkout@v4
       - uses: actions-rs/toolchain@v1
         with:
-          toolchain: nightly-2024-01-04
+          toolchain: nightly-2024-02-06
           override: true
           components: rustfmt, clippy
       - uses: baptiste0928/cargo-install@v1
@@ -172,7 +172,7 @@ jobs:
       - uses: actions/checkout@v4
       - uses: actions-rs/toolchain@v1
         with:
-          toolchain: nightly-2024-01-04
+          toolchain: nightly-2024-02-06
           override: true
           components: rustfmt, clippy
       - uses: baptiste0928/cargo-install@v1
@@ -184,12 +184,12 @@ jobs:
 
   wasm32-tests:
     runs-on: ubuntu-latest
-    # needs: [build,  library-tests, docs]
+    needs: [build, library-tests, docs, python-tests, python-integration-tests]
     steps:
       - uses: actions/checkout@v4
       - uses: actions-rs/toolchain@v1
         with:
-          toolchain: nightly-2024-01-04
+          toolchain: nightly-2024-02-06
           override: true
           components: rustfmt, clippy
       - uses: jetli/wasm-pack-action@v0.4.0
@@ -199,7 +199,7 @@ jobs:
       - name: Install wasm32-unknown-unknown
         run: rustup target add wasm32-unknown-unknown
       - name: Add rust-src
-        run: rustup component add rust-src --toolchain nightly-2024-01-04-x86_64-unknown-linux-gnu
+        run: rustup component add rust-src --toolchain nightly-2024-02-06-x86_64-unknown-linux-gnu
       - name: Run wasm verifier tests
         # on mac:
         #   AR=/opt/homebrew/opt/llvm/bin/llvm-ar CC=/opt/homebrew/opt/llvm/bin/clang wasm-pack test --firefox --headless -- -Z build-std="panic_abort,std" --features web
@@ -207,12 +207,12 @@ jobs:
 
   tutorial:
     runs-on: ubuntu-latest
-    needs: [build, library-tests, docs]
+    needs: [build, library-tests, docs, python-tests, python-integration-tests]
     steps:
       - uses: actions/checkout@v4
       - uses: actions-rs/toolchain@v1
         with:
-          toolchain: nightly-2024-01-04
+          toolchain: nightly-2024-02-06
           override: true
           components: rustfmt, clippy
       - uses: baptiste0928/cargo-install@v1
@@ -224,12 +224,12 @@ jobs:
 
   mock-proving-tests:
     runs-on: non-gpu
-    # needs: [build,  library-tests, docs]
+    needs: [build, library-tests, docs, python-tests, python-integration-tests]
     steps:
       - uses: actions/checkout@v4
       - uses: actions-rs/toolchain@v1
         with:
-          toolchain: nightly-2024-01-04
+          toolchain: nightly-2024-02-06
           override: true
           components: rustfmt, clippy
       - uses: baptiste0928/cargo-install@v1
@@ -281,12 +281,12 @@ jobs:
 
   prove-and-verify-evm-tests:
     runs-on: non-gpu
-    needs: [build, library-tests]
+    needs: [build, library-tests, docs, python-tests, python-integration-tests]
     steps:
       - uses: actions/checkout@v4
       - uses: actions-rs/toolchain@v1
         with:
-          toolchain: nightly-2024-01-04
+          toolchain: nightly-2024-02-06
           override: true
           components: rustfmt, clippy
       - uses: baptiste0928/cargo-install@v1
@@ -303,6 +303,8 @@ jobs:
         with:
           node-version: "18.12.1"
           cache: "pnpm"
+      - name: "Add rust-src"
+        run: rustup component add rust-src --toolchain nightly-2024-02-06-x86_64-unknown-linux-gnu
       - name: Install dependencies for js tests and in-browser-evm-verifier package
         run: |
           pnpm install --no-frozen-lockfile
@@ -324,7 +326,7 @@ jobs:
       - name: Install solc
         run: (hash svm 2>/dev/null || cargo install svm-rs) && svm install 0.8.20 && solc --version
       - name: Install Anvil
-        run: cargo install --git https://github.com/foundry-rs/foundry --rev b320f350156a0fb15c2eb13dc380deb2367c4474 --profile local --locked anvil --force
+        run: cargo install --git https://github.com/foundry-rs/foundry --rev c2233ec9fe61e0920c61c6d779bc707252852037 --profile local --locked anvil --force
       - name: KZG prove and verify tests (EVM + VK rendered seperately)
         run: cargo nextest run --release --verbose tests_evm::kzg_evm_prove_and_verify_render_seperately_ --test-threads 1
       - name: KZG prove and verify tests (EVM + kzg all)
@@ -352,12 +354,12 @@ jobs:
 
   prove-and-verify-tests:
     runs-on: non-gpu
-    needs: [build, library-tests]
+    needs: [build, library-tests, docs, python-tests, python-integration-tests]
     steps:
       - uses: actions/checkout@v4
       - uses: actions-rs/toolchain@v1
         with:
-          toolchain: nightly-2024-01-04
+          toolchain: nightly-2024-02-06
           override: true
           components: rustfmt, clippy
       - uses: jetli/wasm-pack-action@v0.4.0
@@ -365,7 +367,7 @@ jobs:
         run: rustup target add wasm32-unknown-unknown
 
       - name: Add rust-src
-        run: rustup component add rust-src --toolchain nightly-2024-01-04-x86_64-unknown-linux-gnu
+        run: rustup component add rust-src --toolchain nightly-2024-02-06-x86_64-unknown-linux-gnu
       - uses: actions/checkout@v3
       - name: Use pnpm 8
         uses: pnpm/action-setup@v2
@@ -392,6 +394,12 @@ jobs:
       - name: Replace memory definition in nodejs
         run: |
           sed -i "3s|.*|imports['env'] = {memory: new WebAssembly.Memory({initial:20,maximum:65536,shared:true})}|" tests/wasm/nodejs/ezkl.js
+      - name: KZG prove and verify tests (public outputs)
+        run: cargo nextest run --release --verbose tests::kzg_prove_and_verify_tight_lookup_::t
+      - name: IPA prove and verify tests
+        run: cargo nextest run --release --verbose tests::ipa_prove_and_verify_::t --test-threads 1
+      - name: IPA prove and verify tests (ipa outputs)
+        run: cargo nextest run --release --verbose tests::ipa_prove_and_verify_ipa_output
       - name: KZG prove and verify tests (public outputs + column overflow)
         run: cargo nextest run --release --verbose tests::kzg_prove_and_verify_with_overflow_::w
       - name: KZG prove and verify tests single inner col
@@ -408,8 +416,6 @@ jobs:
         run: cargo nextest run --release --verbose tests::kzg_prove_and_verify_with_overflow_fixed_params_
       - name: KZG prove and verify tests (public outputs)
         run: cargo nextest run --release --verbose tests::kzg_prove_and_verify_::t
-      - name: KZG prove and verify tests (public outputs + column overflow)
-        run: cargo nextest run --release --verbose tests::kzg_prove_and_verify_::t
       - name: KZG prove and verify tests (public inputs)
         run: cargo nextest run --release --verbose tests::kzg_prove_and_verify_public_input
       - name: KZG prove and verify tests (fixed params)
@@ -425,11 +431,11 @@ jobs:
       - uses: actions/checkout@v4
       - uses: actions-rs/toolchain@v1
         with:
-          toolchain: nightly-2024-01-04
+          toolchain: nightly-2024-02-06
           override: true
           components: rustfmt, clippy
       - name: Add rust-src
-        run: rustup component add rust-src --toolchain nightly-2024-01-04-x86_64-unknown-linux-gnu
+        run: rustup component add rust-src --toolchain nightly-2024-02-06-x86_64-unknown-linux-gnu
       - uses: actions/checkout@v3
       - uses: baptiste0928/cargo-install@v1
         with:
@@ -452,44 +458,21 @@ jobs:
       - name: KZG prove and verify tests (hashed outputs)
         run: cargo nextest run --release --verbose tests::kzg_prove_and_verify_hashed --features icicle  --test-threads 1
 
-  fuzz-tests:
-    runs-on: ubuntu-latest-32-cores
-    needs: [build, library-tests, python-tests]
-    steps:
-      - uses: actions/checkout@v4
-      - uses: actions-rs/toolchain@v1
-        with:
-          toolchain: nightly-2024-01-04
-          override: true
-          components: rustfmt, clippy
-      - uses: baptiste0928/cargo-install@v1
-        with:
-          crate: cargo-nextest
-          locked: true
-      - name: Install solc
-        run: (hash svm 2>/dev/null || cargo install svm-rs) && svm install 0.8.20 && solc --version
-      - name: Install Anvil
-        run: cargo install --git https://github.com/foundry-rs/foundry --rev b320f350156a0fb15c2eb13dc380deb2367c4474 --profile local --locked anvil --force
-      - name: fuzz tests (EVM)
-        run: cargo nextest run --release --verbose tests_evm::kzg_evm_fuzz_ --test-threads 2
-      # - name: fuzz tests
-      #   run: cargo nextest run --release --verbose tests::kzg_fuzz_ --test-threads 6
-
   prove-and-verify-mock-aggr-tests:
     runs-on: self-hosted
-    needs: [build, library-tests]
+    needs: [build, library-tests, docs, python-tests, python-integration-tests]
     steps:
       - uses: actions/checkout@v4
       - uses: actions-rs/toolchain@v1
         with:
-          toolchain: nightly-2024-01-04
+          toolchain: nightly-2024-02-06
           override: true
           components: rustfmt, clippy
       - uses: baptiste0928/cargo-install@v1
         with:
           crate: cargo-nextest
           locked: true
-      - name: Mock aggr tests
+      - name: Mock aggr tests (KZG)
         run: cargo nextest run --release --verbose tests_aggr::kzg_aggr_mock_prove_and_verify_  --test-threads 8
 
   prove-and-verify-aggr-tests-gpu:
@@ -500,7 +483,7 @@ jobs:
       - uses: actions/checkout@v4
       - uses: actions-rs/toolchain@v1
         with:
-          toolchain: nightly-2024-01-04
+          toolchain: nightly-2024-02-06
           override: true
           components: rustfmt, clippy
       - uses: baptiste0928/cargo-install@v1
@@ -512,29 +495,29 @@ jobs:
 
   prove-and-verify-aggr-tests:
     runs-on: large-self-hosted
-    needs: [build, library-tests]
+    needs: [build, library-tests, docs, python-tests, python-integration-tests]
     steps:
       - uses: actions/checkout@v4
       - uses: actions-rs/toolchain@v1
         with:
-          toolchain: nightly-2024-01-04
+          toolchain: nightly-2024-02-06
           override: true
           components: rustfmt, clippy
       - uses: baptiste0928/cargo-install@v1
         with:
           crate: cargo-nextest
           locked: true
-      - name: KZG )tests
+      - name: KZG tests
         run: cargo nextest run --release --verbose tests_aggr::kzg_aggr_prove_and_verify_  --test-threads 4 -- --include-ignored
 
   prove-and-verify-aggr-evm-tests:
     runs-on: large-self-hosted
-    needs: [build, library-tests]
+    needs: [build, library-tests, docs, python-tests, python-integration-tests]
     steps:
       - uses: actions/checkout@v4
       - uses: actions-rs/toolchain@v1
         with:
-          toolchain: nightly-2024-01-04
+          toolchain: nightly-2024-02-06
           override: true
           components: rustfmt, clippy
       - uses: baptiste0928/cargo-install@v1
@@ -544,7 +527,7 @@ jobs:
       - name: Install solc
         run: (hash svm 2>/dev/null || cargo install svm-rs) && svm install 0.8.20 && solc --version
       - name: Install Anvil
-        run: cargo install --git https://github.com/foundry-rs/foundry --rev b320f350156a0fb15c2eb13dc380deb2367c4474 --profile local --locked anvil --force
+        run: cargo install --git https://github.com/foundry-rs/foundry --rev c2233ec9fe61e0920c61c6d779bc707252852037 --profile local --locked anvil --force
       - name: KZG prove and verify aggr tests
         run: cargo nextest run  --release --verbose tests_evm::kzg_evm_aggr_prove_and_verify_::t --test-threads 4 -- --include-ignored
 
@@ -555,7 +538,7 @@ jobs:
       - uses: actions/checkout@v4
       - uses: actions-rs/toolchain@v1
         with:
-          toolchain: nightly-2024-01-04
+          toolchain: nightly-2024-02-06
           override: true
           components: rustfmt, clippy
       - uses: baptiste0928/cargo-install@v1
@@ -574,18 +557,20 @@ jobs:
       - uses: actions/checkout@v4
       - uses: actions/setup-python@v4
         with:
-          python-version: "3.7"
+          python-version: "3.12"
       - uses: actions-rs/toolchain@v1
         with:
-          toolchain: nightly-2024-01-04
+          toolchain: nightly-2024-02-06
           override: true
           components: rustfmt, clippy
+      - name: Install cmake
+        run: sudo apt-get install -y cmake
       - name: Install solc
         run: (hash svm 2>/dev/null || cargo install svm-rs) && svm install 0.8.20 && solc --version
       - name: Setup Virtual Env and Install python dependencies
-        run: python -m venv .env; source .env/bin/activate; pip install -r requirements.txt;
+        run: python -m venv .env --clear; source .env/bin/activate; pip install -r requirements.txt;
       - name: Install Anvil
-        run: cargo install --git https://github.com/foundry-rs/foundry --rev b320f350156a0fb15c2eb13dc380deb2367c4474 --profile local --locked anvil --force
+        run: cargo install --git https://github.com/foundry-rs/foundry --rev c2233ec9fe61e0920c61c6d779bc707252852037 --profile local --locked anvil --force
       - name: Build python ezkl
         run: source .env/bin/activate; unset CONDA_PREFIX; maturin develop --features python-bindings --release
       - name: Run pytest
@@ -593,15 +578,15 @@ jobs:
 
   accuracy-measurement-tests:
     runs-on: ubuntu-latest-32-cores
-    # needs: [build, library-tests, docs]
+    needs: [build, library-tests, docs, python-tests, python-integration-tests]
     steps:
       - uses: actions/checkout@v4
       - uses: actions/setup-python@v4
         with:
-          python-version: "3.7"
+          python-version: "3.12"
       - uses: actions-rs/toolchain@v1
         with:
-          toolchain: nightly-2024-01-04
+          toolchain: nightly-2024-02-06
           override: true
           components: rustfmt, clippy
       - uses: baptiste0928/cargo-install@v1
@@ -609,7 +594,7 @@ jobs:
           crate: cargo-nextest
           locked: true
       - name: Setup Virtual Env and Install python dependencies
-        run: python -m venv .env; source .env/bin/activate; pip install -r requirements.txt;
+        run: python -m venv .env --clear; source .env/bin/activate; pip install -r requirements.txt;
       - name: Build python ezkl
         run: source .env/bin/activate; unset CONDA_PREFIX; maturin develop --features python-bindings --release
       - name: Div rebase
@@ -629,10 +614,10 @@ jobs:
       - uses: actions/checkout@v4
       - uses: actions/setup-python@v4
         with:
-          python-version: "3.9"
+          python-version: "3.11"
       - uses: actions-rs/toolchain@v1
         with:
-          toolchain: nightly-2024-01-04
+          toolchain: nightly-2024-02-06
           override: true
           components: rustfmt, clippy
       - uses: baptiste0928/cargo-install@v1
@@ -642,11 +627,15 @@ jobs:
       - name: Install solc
         run: (hash svm 2>/dev/null || cargo install svm-rs) && svm install 0.8.20 && solc --version
       - name: Install Anvil
-        run: cargo install --git https://github.com/foundry-rs/foundry --rev b320f350156a0fb15c2eb13dc380deb2367c4474 --profile local --locked anvil --force
+        run: cargo install --git https://github.com/foundry-rs/foundry --rev c2233ec9fe61e0920c61c6d779bc707252852037 --profile local --locked anvil --force
+      - name: Install pip
+        run: python -m ensurepip --upgrade
       - name: Setup Virtual Env and Install python dependencies
-        run: python -m venv .env; source .env/bin/activate; pip install -r requirements.txt;
+        run: python -m venv .env --clear; source .env/bin/activate; pip install -r requirements.txt; python -m ensurepip --upgrade
       - name: Build python ezkl
         run: source .env/bin/activate; unset CONDA_PREFIX; maturin develop --features python-bindings --release
+      - name: Tictactoe tutorials
+        run: source .env/bin/activate; cargo nextest run py_tests::tests::tictactoe_ --test-threads 1
       # - name: authenticate-kaggle-cli
       #   shell: bash
       #   env:
@@ -657,12 +646,10 @@ jobs:
       #     echo $KAGGLE_API_KEY > /home/ubuntu/.kaggle/kaggle.json
       #     chmod 600 /home/ubuntu/.kaggle/kaggle.json
       - name: All notebooks
-        run: source .env/bin/activate; cargo nextest run py_tests::tests::run_notebook_ --no-capture
+        run: source .env/bin/activate; cargo nextest run py_tests::tests::run_notebook_ --test-threads 1
       - name: Voice tutorial
         run: source .env/bin/activate; cargo nextest run py_tests::tests::voice_
       - name: NBEATS tutorial
         run: source .env/bin/activate; cargo nextest run py_tests::tests::nbeats_
-      - name: Tictactoe tutorials
-        run: source .env/bin/activate; cargo nextest run py_tests::tests::tictactoe_ --no-capture
       # - name: Postgres tutorials
       #   run: source .env/bin/activate; cargo nextest run py_tests::tests::postgres_ --test-threads 1

.github/workflows/tagging.yml

@@ -14,6 +14,40 @@ jobs:
       - uses: actions/checkout@v4
       - name: Bump version and push tag
         id: tag_version
-        uses: mathieudutour/github-tag-action@v6.1
+        uses: mathieudutour/github-tag-action@v6.2
         with:
           github_token: ${{ secrets.GITHUB_TOKEN }}
+
+      - name: Set Cargo.toml version to match github tag for docs
+        shell: bash
+        env:
+          RELEASE_TAG: ${{ steps.tag_version.outputs.new_tag }}
+        run: |
+          mv docs/python/src/conf.py docs/python/src/conf.py.orig
+          sed "s/0\\.0\\.0/${RELEASE_TAG//v}/" docs/python/src/conf.py.orig >docs/python/src/conf.py
+          rm docs/python/src/conf.py.orig
+          mv docs/python/requirements-docs.txt docs/python/requirements-docs.txt.orig
+          sed "s/0\\.0\\.0/${RELEASE_TAG//v}/" docs/python/requirements-docs.txt.orig >docs/python/requirements-docs.txt
+          rm docs/python/requirements-docs.txt.orig
+
+      - name: Commit files and create tag
+        env:
+          RELEASE_TAG: ${{ steps.tag_version.outputs.new_tag }}
+        run: |
+          git config --local user.email "github-actions[bot]@users.noreply.github.com"
+          git config --local user.name "github-actions[bot]"
+          git fetch --tags
+          git checkout -b release-$RELEASE_TAG
+          git add .
+          git commit -m "ci: update version string in docs"
+          git tag -d $RELEASE_TAG
+          git tag $RELEASE_TAG
+
+      - name: Push changes
+        uses: ad-m/github-push-action@master
+        env:
+          RELEASE_TAG: ${{ steps.tag_version.outputs.new_tag }}
+        with:
+          branch: release-${{ steps.tag_version.outputs.new_tag }}
+          force: true
+          tags: true

.gitignore

@@ -48,4 +48,6 @@ node_modules
 /dist
 timingData.json
 !tests/wasm/pk.key
-!tests/wasm/vk.key
+!tests/wasm/vk.key
+docs/python/build
+!tests/wasm/vk_aggr.key

.python-version

@@ -0,0 +1 @@
+3.12.1

.readthedocs.yaml

@@ -0,0 +1,26 @@
+# .readthedocs.yaml
+# Read the Docs configuration file
+# See https://docs.readthedocs.io/en/stable/config-file/v2.html for details
+
+version: 2
+
+build:
+  os: ubuntu-22.04
+  tools:
+    python: "3.12"
+
+# Build documentation in the "docs/" directory with Sphinx
+sphinx:
+  configuration: ./docs/python/src/conf.py
+
+# Optionally build your docs in additional formats such as PDF and ePub
+# formats:
+#    - pdf
+#    - epub
+
+# Optional but recommended, declare the Python requirements required
+# to build your documentation
+# See https://docs.readthedocs.io/en/stable/guides/reproducible-builds.html
+python:
+  install:
+  - requirements: ./docs/python/requirements-docs.txt

Cargo.toml

@@ -15,14 +15,14 @@ crate-type = ["cdylib", "rlib"]
 
 
 [dependencies]
-halo2_gadgets = { git = "https://github.com/zkonduit/halo2", branch = "main" }
-halo2_proofs = { git = "https://github.com/zkonduit/halo2", branch = "main" }
+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 = [
     "derive_serde",
 ] }
 rand = { version = "0.8", default_features = false }
 itertools = { version = "0.10.3", default_features = false }
-clap = { version = "4.3.3", features = ["derive"] }
+clap = { version = "4.5.3", features = ["derive"] }
 serde = { version = "1.0.126", features = ["derive"], optional = true }
 serde_json = { version = "1.0.97", default_features = false, features = [
     "float_roundtrip",
@@ -80,7 +80,7 @@ pyo3-asyncio = { version = "0.20.0", features = [
     "tokio-runtime",
 ], default_features = false, optional = true }
 pyo3-log = { version = "0.9.0", default_features = false, optional = true }
-tract-onnx = { git = "https://github.com/sonos/tract/", rev = "7b1aa33b2f7d1f19b80e270c83320f0f94daff69", default_features = false, optional = true }
+tract-onnx = { git = "https://github.com/sonos/tract/", rev = "681a096f02c9d7d363102d9fb0e446d1710ac2c8", default_features = false, optional = true }
 tabled = { version = "0.12.0", optional = true }
 
 
@@ -95,10 +95,10 @@ getrandom = { version = "0.2.8", features = ["js"] }
 instant = { version = "0.1", features = ["wasm-bindgen", "inaccurate"] }
 
 [target.'cfg(all(target_arch = "wasm32", target_os = "unknown"))'.dependencies]
-wasm-bindgen-rayon = { version = "1.0", optional = true }
-wasm-bindgen-test = "0.3.34"
-serde-wasm-bindgen = "0.4"
-wasm-bindgen = { version = "0.2.81", features = ["serde-serialize"] }
+wasm-bindgen-rayon = { version = "1.2.1", optional = true }
+wasm-bindgen-test = "0.3.42"
+serde-wasm-bindgen = "0.6.5"
+wasm-bindgen = { version = "0.2.92", features = ["serde-serialize"] }
 console_error_panic_hook = "0.1.7"
 wasm-bindgen-console-logger = "0.1.1"
 
@@ -203,5 +203,9 @@ no-banner = []
 [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" }
 
+[patch.'https://github.com/zkonduit/halo2']
+halo2_proofs = { git = "https://github.com/zkonduit/halo2?branch=ac/optional-selector-poly#54f54453cf186aa5d89579c4e7663f9a27cfb89a", package = "halo2_proofs", branch = "ac/optional-selector-poly" }
+
+
 [profile.release]
 rustflags = ["-C", "relocation-model=pic"]

README.md

@@ -31,9 +31,9 @@ EZKL
 
 [![Notebook](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/github/zkonduit/ezkl/blob/main/examples/notebooks/simple_demo_all_public.ipynb) 
 
-In the backend we use [Halo2](https://github.com/privacy-scaling-explorations/halo2) as a proof system.
+In the backend we use the collaboratively-developed [Halo2](https://github.com/privacy-scaling-explorations/halo2) as a proof system.
 
-The generated proofs can then be used on-chain to verify computation, only the Ethereum Virtual Machine (EVM) is supported at the moment.
+The generated proofs can then be verified with much less computational resources, including on-chain (with the Ethereum Virtual Machine), in a browser, or on a device. 
 
 - If you have any questions, we'd love for you to open up a discussion topic in [Discussions](https://github.com/zkonduit/ezkl/discussions). Alternatively, you can join the ✨[EZKL Community Telegram Group](https://t.me/+QRzaRvTPIthlYWMx)💫.
 
@@ -45,6 +45,8 @@ The generated proofs can then be used on-chain to verify computation, only the E
 
 ### getting started ⚙️
 
+The easiest way to get started is to try out a notebook. 
+
 #### Python
 Install the python bindings by calling.
 
@@ -70,7 +72,7 @@ curl https://raw.githubusercontent.com/zkonduit/ezkl/main/install_ezkl_cli.sh |
 
 https://user-images.githubusercontent.com/45801863/236771676-5bbbbfd1-ba6f-418a-902e-20738ce0e9f0.mp4
 
-For more details visit the [docs](https://docs.ezkl.xyz).
+For more details visit the [docs](https://docs.ezkl.xyz). The CLI is faster than Python, as it has less overhead. For even more speed and convenience, check out the [remote proving service](https://ei40vx5x6j0.typeform.com/to/sFv1oxvb), which feels like the CLI but is backed by a tuned cluster.
 
 Build the auto-generated rust documentation and open the docs in your browser locally. `cargo doc --open`
 
@@ -124,17 +126,6 @@ unset ENABLE_ICICLE_GPU
 
 **NOTE:** Even with the above environment variable set, icicle is disabled for circuits where k <= 8. To change the value of `k` where icicle is enabled, you can set the environment variable `ICICLE_SMALL_K`.
 
-### repos
-
-The EZKL project has several libraries and repos. 
-
-| Repo | Description |
-| --- | --- |
-| [@zkonduit/ezkl](https://github.com/zkonduit/ezkl) | the main ezkl repo in rust with wasm and python bindings |
-| [@zkonduit/ezkljs](https://github.com/zkonduit/ezkljs) | typescript and javascript tooling to help integrate ezkl into web apps |
-
-----------------------
-
 ### contributing 🌎
 
 If you're interested in contributing and are unsure where to start, reach out to one of the maintainers:
@@ -151,7 +142,7 @@ More broadly:
 - To report bugs or request new features [create a new issue within Issues](https://github.com/zkonduit/ezkl/issues) to inform the greater community.
 
 
-Unless you explicitly state otherwise, any contribution intentionally submitted for inclusion in the work by you shall be licensed to Zkonduit Inc. under the terms and conditions specified in the [CLA](https://github.com/zkonduit/ezkl/blob/main/cla.md), which you agree to by intentionally submitting a contribution. In particular, you have the right to submit the contribution and we can distribute it under the Apache 2.0 license, among other terms and conditions. 
+Any contribution intentionally submitted for inclusion in the work by you shall be licensed to Zkonduit Inc. under the terms and conditions specified in the [CLA](https://github.com/zkonduit/ezkl/blob/main/cla.md), which you agree to by intentionally submitting a contribution. In particular, you have the right to submit the contribution and we can distribute it, among other terms and conditions. 
 
 ### no security guarantees
 
@@ -159,4 +150,7 @@ 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. 
 
+### 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,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.

benches/accum_conv.rs

@@ -2,11 +2,13 @@ use criterion::{criterion_group, criterion_main, BenchmarkId, Criterion, Through
 use ezkl::circuit::poly::PolyOp;
 use ezkl::circuit::*;
 use ezkl::pfsys::create_keys;
-use ezkl::pfsys::create_proof_circuit_kzg;
+use ezkl::pfsys::create_proof_circuit;
 use ezkl::pfsys::srs::gen_srs;
 use ezkl::pfsys::TranscriptType;
 use ezkl::tensor::*;
 use halo2_proofs::poly::kzg::commitment::KZGCommitmentScheme;
+use halo2_proofs::poly::kzg::multiopen::ProverSHPLONK;
+use halo2_proofs::poly::kzg::multiopen::VerifierSHPLONK;
 use halo2_proofs::poly::kzg::strategy::SingleStrategy;
 use halo2_proofs::{
     arithmetic::Field,
@@ -15,6 +17,7 @@ use halo2_proofs::{
 };
 use halo2curves::bn256::{Bn256, Fr};
 use rand::rngs::OsRng;
+use snark_verifier::system::halo2::transcript::evm::EvmTranscript;
 
 static mut KERNEL_HEIGHT: usize = 2;
 static mut KERNEL_WIDTH: usize = 2;
@@ -67,8 +70,8 @@ impl Circuit<Fr> for MyCircuit {
                         &mut region,
                         &[self.image.clone(), self.kernel.clone(), self.bias.clone()],
                         Box::new(PolyOp::Conv {
-                            padding: [(0, 0); 2],
-                            stride: (1, 1),
+                            padding: vec![(0, 0)],
+                            stride: vec![1; 2],
                         }),
                     )
                     .unwrap();
@@ -121,28 +124,35 @@ fn runcnvrl(c: &mut Criterion) {
             group.throughput(Throughput::Elements(*size as u64));
             group.bench_with_input(BenchmarkId::new("pk", size), &size, |b, &_| {
                 b.iter(|| {
-                    create_keys::<KZGCommitmentScheme<Bn256>, Fr, MyCircuit>(
-                        &circuit, &params, true,
-                    )
-                    .unwrap();
+                    create_keys::<KZGCommitmentScheme<Bn256>, MyCircuit>(&circuit, &params, true)
+                        .unwrap();
                 });
             });
 
-            let pk =
-                create_keys::<KZGCommitmentScheme<Bn256>, Fr, MyCircuit>(&circuit, &params, true)
-                    .unwrap();
+            let pk = create_keys::<KZGCommitmentScheme<Bn256>, MyCircuit>(&circuit, &params, true)
+                .unwrap();
 
             group.throughput(Throughput::Elements(*size as u64));
             group.bench_with_input(BenchmarkId::new("prove", size), &size, |b, &_| {
                 b.iter(|| {
-                    let prover = create_proof_circuit_kzg(
+                    let prover = create_proof_circuit::<
+                        KZGCommitmentScheme<_>,
+                        MyCircuit,
+                        ProverSHPLONK<_>,
+                        VerifierSHPLONK<_>,
+                        SingleStrategy<_>,
+                        _,
+                        EvmTranscript<_, _, _, _>,
+                        EvmTranscript<_, _, _, _>,
+                    >(
                         circuit.clone(),
+                        vec![],
                         &params,
-                        None,
                         &pk,
-                        TranscriptType::EVM,
-                        SingleStrategy::new(&params),
                         CheckMode::UNSAFE,
+                        ezkl::Commitments::KZG,
+                        TranscriptType::EVM,
+                        None,
                         None,
                     );
                     prover.unwrap();

benches/accum_dot.rs

@@ -1,11 +1,13 @@
 use criterion::{criterion_group, criterion_main, BenchmarkId, Criterion, Throughput};
 use ezkl::circuit::poly::PolyOp;
 use ezkl::circuit::*;
-use ezkl::pfsys::create_proof_circuit_kzg;
+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;
+use halo2_proofs::poly::kzg::multiopen::VerifierSHPLONK;
 use halo2_proofs::poly::kzg::strategy::SingleStrategy;
 use halo2_proofs::{
     arithmetic::Field,
@@ -14,6 +16,7 @@ use halo2_proofs::{
 };
 use halo2curves::bn256::{Bn256, Fr};
 use rand::rngs::OsRng;
+use snark_verifier::system::halo2::transcript::evm::EvmTranscript;
 use std::marker::PhantomData;
 
 static mut LEN: usize = 4;
@@ -90,25 +93,35 @@ fn rundot(c: &mut Criterion) {
         group.throughput(Throughput::Elements(len as u64));
         group.bench_with_input(BenchmarkId::new("pk", len), &len, |b, &_| {
             b.iter(|| {
-                create_keys::<KZGCommitmentScheme<Bn256>, Fr, MyCircuit>(&circuit, &params, true)
+                create_keys::<KZGCommitmentScheme<Bn256>, MyCircuit>(&circuit, &params, true)
                     .unwrap();
             });
         });
 
-        let pk = create_keys::<KZGCommitmentScheme<Bn256>, Fr, MyCircuit>(&circuit, &params, true)
-            .unwrap();
+        let pk =
+            create_keys::<KZGCommitmentScheme<Bn256>, MyCircuit>(&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_kzg(
+                let prover = create_proof_circuit::<
+                    KZGCommitmentScheme<_>,
+                    MyCircuit,
+                    ProverSHPLONK<_>,
+                    VerifierSHPLONK<_>,
+                    SingleStrategy<_>,
+                    _,
+                    EvmTranscript<_, _, _, _>,
+                    EvmTranscript<_, _, _, _>,
+                >(
                     circuit.clone(),
+                    vec![],
                     &params,
-                    None,
                     &pk,
-                    TranscriptType::EVM,
-                    SingleStrategy::new(&params),
                     CheckMode::UNSAFE,
+                    ezkl::Commitments::KZG,
+                    TranscriptType::EVM,
+                    None,
                     None,
                 );
                 prover.unwrap();

benches/accum_einsum_matmul.rs

@@ -1,11 +1,13 @@
 use criterion::{criterion_group, criterion_main, BenchmarkId, Criterion, Throughput};
 use ezkl::circuit::poly::PolyOp;
 use ezkl::circuit::*;
-use ezkl::pfsys::create_proof_circuit_kzg;
+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;
+use halo2_proofs::poly::kzg::multiopen::VerifierSHPLONK;
 use halo2_proofs::poly::kzg::strategy::SingleStrategy;
 use halo2_proofs::{
     arithmetic::Field,
@@ -14,6 +16,7 @@ use halo2_proofs::{
 };
 use halo2curves::bn256::{Bn256, Fr};
 use rand::rngs::OsRng;
+use snark_verifier::system::halo2::transcript::evm::EvmTranscript;
 use std::marker::PhantomData;
 
 static mut LEN: usize = 4;
@@ -94,25 +97,35 @@ fn runmatmul(c: &mut Criterion) {
         group.throughput(Throughput::Elements(len as u64));
         group.bench_with_input(BenchmarkId::new("pk", len), &len, |b, &_| {
             b.iter(|| {
-                create_keys::<KZGCommitmentScheme<Bn256>, Fr, MyCircuit>(&circuit, &params, true)
+                create_keys::<KZGCommitmentScheme<Bn256>, MyCircuit>(&circuit, &params, true)
                     .unwrap();
             });
         });
 
-        let pk = create_keys::<KZGCommitmentScheme<Bn256>, Fr, MyCircuit>(&circuit, &params, true)
-            .unwrap();
+        let pk =
+            create_keys::<KZGCommitmentScheme<Bn256>, MyCircuit>(&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_kzg(
+                let prover = create_proof_circuit::<
+                    KZGCommitmentScheme<_>,
+                    MyCircuit,
+                    ProverSHPLONK<_>,
+                    VerifierSHPLONK<_>,
+                    SingleStrategy<_>,
+                    _,
+                    EvmTranscript<_, _, _, _>,
+                    EvmTranscript<_, _, _, _>,
+                >(
                     circuit.clone(),
+                    vec![],
                     &params,
-                    None,
                     &pk,
-                    TranscriptType::EVM,
-                    SingleStrategy::new(&params),
                     CheckMode::UNSAFE,
+                    ezkl::Commitments::KZG,
+                    TranscriptType::EVM,
+                    None,
                     None,
                 );
                 prover.unwrap();

benches/accum_matmul_relu.rs

@@ -4,17 +4,20 @@ use ezkl::circuit::*;
 
 use ezkl::circuit::lookup::LookupOp;
 use ezkl::circuit::poly::PolyOp;
-use ezkl::pfsys::create_proof_circuit_kzg;
+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;
+use halo2_proofs::poly::kzg::multiopen::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 snark_verifier::system::halo2::transcript::evm::EvmTranscript;
 use std::marker::PhantomData;
 
 const BITS: Range = (-32768, 32768);
@@ -112,25 +115,35 @@ fn runmatmul(c: &mut Criterion) {
         group.throughput(Throughput::Elements(len as u64));
         group.bench_with_input(BenchmarkId::new("pk", len), &len, |b, &_| {
             b.iter(|| {
-                create_keys::<KZGCommitmentScheme<Bn256>, Fr, MyCircuit>(&circuit, &params, true)
+                create_keys::<KZGCommitmentScheme<Bn256>, MyCircuit>(&circuit, &params, true)
                     .unwrap();
             });
         });
 
-        let pk = create_keys::<KZGCommitmentScheme<Bn256>, Fr, MyCircuit>(&circuit, &params, true)
-            .unwrap();
+        let pk =
+            create_keys::<KZGCommitmentScheme<Bn256>, MyCircuit>(&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_kzg(
+                let prover = create_proof_circuit::<
+                    KZGCommitmentScheme<_>,
+                    MyCircuit,
+                    ProverSHPLONK<_>,
+                    VerifierSHPLONK<_>,
+                    SingleStrategy<_>,
+                    _,
+                    EvmTranscript<_, _, _, _>,
+                    EvmTranscript<_, _, _, _>,
+                >(
                     circuit.clone(),
+                    vec![],
                     &params,
-                    None,
                     &pk,
+                    CheckMode::UNSAFE,
+                    ezkl::Commitments::KZG,
                     TranscriptType::EVM,
-                    SingleStrategy::new(&params),
-                    CheckMode::SAFE,
+                    None,
                     None,
                 );
                 prover.unwrap();

benches/accum_matmul_relu_overflow.rs

@@ -4,17 +4,20 @@ use ezkl::circuit::*;
 use ezkl::circuit::lookup::LookupOp;
 use ezkl::circuit::poly::PolyOp;
 use ezkl::circuit::table::Range;
-use ezkl::pfsys::create_proof_circuit_kzg;
+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;
+use halo2_proofs::poly::kzg::multiopen::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 snark_verifier::system::halo2::transcript::evm::EvmTranscript;
 use std::marker::PhantomData;
 
 const BITS: Range = (-8180, 8180);
@@ -115,25 +118,35 @@ fn runmatmul(c: &mut Criterion) {
         group.throughput(Throughput::Elements(k as u64));
         group.bench_with_input(BenchmarkId::new("pk", k), &k, |b, &_| {
             b.iter(|| {
-                create_keys::<KZGCommitmentScheme<Bn256>, Fr, MyCircuit>(&circuit, &params, true)
+                create_keys::<KZGCommitmentScheme<Bn256>, MyCircuit>(&circuit, &params, true)
                     .unwrap();
             });
         });
 
-        let pk = create_keys::<KZGCommitmentScheme<Bn256>, Fr, MyCircuit>(&circuit, &params, true)
-            .unwrap();
+        let pk =
+            create_keys::<KZGCommitmentScheme<Bn256>, MyCircuit>(&circuit, &params, true).unwrap();
 
         group.throughput(Throughput::Elements(k as u64));
         group.bench_with_input(BenchmarkId::new("prove", k), &k, |b, &_| {
             b.iter(|| {
-                let prover = create_proof_circuit_kzg(
+                let prover = create_proof_circuit::<
+                    KZGCommitmentScheme<_>,
+                    MyCircuit,
+                    ProverSHPLONK<_>,
+                    VerifierSHPLONK<_>,
+                    SingleStrategy<_>,
+                    _,
+                    EvmTranscript<_, _, _, _>,
+                    EvmTranscript<_, _, _, _>,
+                >(
                     circuit.clone(),
+                    vec![],
                     &params,
-                    None,
                     &pk,
+                    CheckMode::UNSAFE,
+                    ezkl::Commitments::KZG,
                     TranscriptType::EVM,
-                    SingleStrategy::new(&params),
-                    CheckMode::SAFE,
+                    None,
                     None,
                 );
                 prover.unwrap();

benches/accum_sum.rs

@@ -1,11 +1,13 @@
 use criterion::{criterion_group, criterion_main, BenchmarkId, Criterion, Throughput};
 use ezkl::circuit::poly::PolyOp;
 use ezkl::circuit::*;
-use ezkl::pfsys::create_proof_circuit_kzg;
+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;
+use halo2_proofs::poly::kzg::multiopen::VerifierSHPLONK;
 use halo2_proofs::poly::kzg::strategy::SingleStrategy;
 use halo2_proofs::{
     arithmetic::Field,
@@ -14,6 +16,7 @@ use halo2_proofs::{
 };
 use halo2curves::bn256::{Bn256, Fr};
 use rand::rngs::OsRng;
+use snark_verifier::system::halo2::transcript::evm::EvmTranscript;
 use std::marker::PhantomData;
 
 static mut LEN: usize = 4;
@@ -86,25 +89,35 @@ fn runsum(c: &mut Criterion) {
         group.throughput(Throughput::Elements(len as u64));
         group.bench_with_input(BenchmarkId::new("pk", len), &len, |b, &_| {
             b.iter(|| {
-                create_keys::<KZGCommitmentScheme<Bn256>, Fr, MyCircuit>(&circuit, &params, true)
+                create_keys::<KZGCommitmentScheme<Bn256>, MyCircuit>(&circuit, &params, true)
                     .unwrap();
             });
         });
 
-        let pk = create_keys::<KZGCommitmentScheme<Bn256>, Fr, MyCircuit>(&circuit, &params, true)
-            .unwrap();
+        let pk =
+            create_keys::<KZGCommitmentScheme<Bn256>, MyCircuit>(&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_kzg(
+                let prover = create_proof_circuit::<
+                    KZGCommitmentScheme<_>,
+                    MyCircuit,
+                    ProverSHPLONK<_>,
+                    VerifierSHPLONK<_>,
+                    SingleStrategy<_>,
+                    _,
+                    EvmTranscript<_, _, _, _>,
+                    EvmTranscript<_, _, _, _>,
+                >(
                     circuit.clone(),
+                    vec![],
                     &params,
-                    None,
                     &pk,
-                    TranscriptType::EVM,
-                    SingleStrategy::new(&params),
                     CheckMode::UNSAFE,
+                    ezkl::Commitments::KZG,
+                    TranscriptType::EVM,
+                    None,
                     None,
                 );
                 prover.unwrap();

benches/accum_sumpool.rs

@@ -2,11 +2,13 @@ use criterion::{criterion_group, criterion_main, BenchmarkId, Criterion, Through
 use ezkl::circuit::hybrid::HybridOp;
 use ezkl::circuit::*;
 use ezkl::pfsys::create_keys;
-use ezkl::pfsys::create_proof_circuit_kzg;
+use ezkl::pfsys::create_proof_circuit;
 use ezkl::pfsys::srs::gen_srs;
 use ezkl::pfsys::TranscriptType;
 use ezkl::tensor::*;
 use halo2_proofs::poly::kzg::commitment::KZGCommitmentScheme;
+use halo2_proofs::poly::kzg::multiopen::ProverSHPLONK;
+use halo2_proofs::poly::kzg::multiopen::VerifierSHPLONK;
 use halo2_proofs::poly::kzg::strategy::SingleStrategy;
 use halo2_proofs::{
     arithmetic::Field,
@@ -15,6 +17,7 @@ use halo2_proofs::{
 };
 use halo2curves::bn256::{Bn256, Fr};
 use rand::rngs::OsRng;
+use snark_verifier::system::halo2::transcript::evm::EvmTranscript;
 
 static mut IMAGE_HEIGHT: usize = 2;
 static mut IMAGE_WIDTH: usize = 2;
@@ -62,9 +65,9 @@ impl Circuit<Fr> for MyCircuit {
                         &mut region,
                         &[self.image.clone()],
                         Box::new(HybridOp::SumPool {
-                            padding: [(0, 0); 2],
-                            stride: (1, 1),
-                            kernel_shape: (2, 2),
+                            padding: vec![(0, 0); 2],
+                            stride: vec![1, 1],
+                            kernel_shape: vec![2, 2],
                             normalized: false,
                         }),
                     )
@@ -101,28 +104,35 @@ fn runsumpool(c: &mut Criterion) {
             group.throughput(Throughput::Elements(*size as u64));
             group.bench_with_input(BenchmarkId::new("pk", size), &size, |b, &_| {
                 b.iter(|| {
-                    create_keys::<KZGCommitmentScheme<Bn256>, Fr, MyCircuit>(
-                        &circuit, &params, true,
-                    )
-                    .unwrap();
+                    create_keys::<KZGCommitmentScheme<Bn256>, MyCircuit>(&circuit, &params, true)
+                        .unwrap();
                 });
             });
 
-            let pk =
-                create_keys::<KZGCommitmentScheme<Bn256>, Fr, MyCircuit>(&circuit, &params, true)
-                    .unwrap();
+            let pk = create_keys::<KZGCommitmentScheme<Bn256>, MyCircuit>(&circuit, &params, true)
+                .unwrap();
 
             group.throughput(Throughput::Elements(*size as u64));
             group.bench_with_input(BenchmarkId::new("prove", size), &size, |b, &_| {
                 b.iter(|| {
-                    let prover = create_proof_circuit_kzg(
+                    let prover = create_proof_circuit::<
+                        KZGCommitmentScheme<_>,
+                        MyCircuit,
+                        ProverSHPLONK<_>,
+                        VerifierSHPLONK<_>,
+                        SingleStrategy<_>,
+                        _,
+                        EvmTranscript<_, _, _, _>,
+                        EvmTranscript<_, _, _, _>,
+                    >(
                         circuit.clone(),
+                        vec![],
                         &params,
-                        None,
                         &pk,
-                        TranscriptType::EVM,
-                        SingleStrategy::new(&params),
                         CheckMode::UNSAFE,
+                        ezkl::Commitments::KZG,
+                        TranscriptType::EVM,
+                        None,
                         None,
                     );
                     prover.unwrap();

benches/pairwise_add.rs

@@ -1,11 +1,13 @@
 use criterion::{criterion_group, criterion_main, BenchmarkId, Criterion, Throughput};
 use ezkl::circuit::poly::PolyOp;
 use ezkl::circuit::*;
-use ezkl::pfsys::create_proof_circuit_kzg;
+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;
+use halo2_proofs::poly::kzg::multiopen::VerifierSHPLONK;
 use halo2_proofs::poly::kzg::strategy::SingleStrategy;
 use halo2_proofs::{
     arithmetic::Field,
@@ -14,6 +16,7 @@ use halo2_proofs::{
 };
 use halo2curves::bn256::{Bn256, Fr};
 use rand::rngs::OsRng;
+use snark_verifier::system::halo2::transcript::evm::EvmTranscript;
 use std::marker::PhantomData;
 
 static mut LEN: usize = 4;
@@ -84,25 +87,35 @@ fn runadd(c: &mut Criterion) {
         group.throughput(Throughput::Elements(len as u64));
         group.bench_with_input(BenchmarkId::new("pk", len), &len, |b, &_| {
             b.iter(|| {
-                create_keys::<KZGCommitmentScheme<Bn256>, Fr, MyCircuit>(&circuit, &params, true)
+                create_keys::<KZGCommitmentScheme<Bn256>, MyCircuit>(&circuit, &params, true)
                     .unwrap();
             });
         });
 
-        let pk = create_keys::<KZGCommitmentScheme<Bn256>, Fr, MyCircuit>(&circuit, &params, true)
-            .unwrap();
+        let pk =
+            create_keys::<KZGCommitmentScheme<Bn256>, MyCircuit>(&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_kzg(
+                let prover = create_proof_circuit::<
+                    KZGCommitmentScheme<_>,
+                    MyCircuit,
+                    ProverSHPLONK<_>,
+                    VerifierSHPLONK<_>,
+                    SingleStrategy<_>,
+                    _,
+                    EvmTranscript<_, _, _, _>,
+                    EvmTranscript<_, _, _, _>,
+                >(
                     circuit.clone(),
+                    vec![],
                     &params,
-                    None,
                     &pk,
+                    CheckMode::UNSAFE,
+                    ezkl::Commitments::KZG,
                     TranscriptType::EVM,
-                    SingleStrategy::new(&params),
-                    CheckMode::SAFE,
+                    None,
                     None,
                 );
                 prover.unwrap();

benches/pairwise_pow.rs

@@ -2,11 +2,13 @@ use criterion::{criterion_group, criterion_main, BenchmarkId, Criterion, Through
 use ezkl::circuit::poly::PolyOp;
 use ezkl::circuit::region::RegionCtx;
 use ezkl::circuit::*;
-use ezkl::pfsys::create_proof_circuit_kzg;
+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;
+use halo2_proofs::poly::kzg::multiopen::VerifierSHPLONK;
 use halo2_proofs::poly::kzg::strategy::SingleStrategy;
 use halo2_proofs::{
     arithmetic::Field,
@@ -15,6 +17,7 @@ use halo2_proofs::{
 };
 use halo2curves::bn256::{Bn256, Fr};
 use rand::rngs::OsRng;
+use snark_verifier::system::halo2::transcript::evm::EvmTranscript;
 use std::marker::PhantomData;
 
 static mut LEN: usize = 4;
@@ -83,25 +86,35 @@ fn runpow(c: &mut Criterion) {
         group.throughput(Throughput::Elements(len as u64));
         group.bench_with_input(BenchmarkId::new("pk", len), &len, |b, &_| {
             b.iter(|| {
-                create_keys::<KZGCommitmentScheme<Bn256>, Fr, MyCircuit>(&circuit, &params, true)
+                create_keys::<KZGCommitmentScheme<Bn256>, MyCircuit>(&circuit, &params, true)
                     .unwrap();
             });
         });
 
-        let pk = create_keys::<KZGCommitmentScheme<Bn256>, Fr, MyCircuit>(&circuit, &params, true)
-            .unwrap();
+        let pk =
+            create_keys::<KZGCommitmentScheme<Bn256>, MyCircuit>(&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_kzg(
+                let prover = create_proof_circuit::<
+                    KZGCommitmentScheme<_>,
+                    MyCircuit,
+                    ProverSHPLONK<_>,
+                    VerifierSHPLONK<_>,
+                    SingleStrategy<_>,
+                    _,
+                    EvmTranscript<_, _, _, _>,
+                    EvmTranscript<_, _, _, _>,
+                >(
                     circuit.clone(),
+                    vec![],
                     &params,
-                    None,
                     &pk,
+                    CheckMode::UNSAFE,
+                    ezkl::Commitments::KZG,
                     TranscriptType::EVM,
-                    SingleStrategy::new(&params),
-                    CheckMode::SAFE,
+                    None,
                     None,
                 );
                 prover.unwrap();

benches/poseidon.rs

@@ -1,15 +1,18 @@
+use std::collections::HashMap;
+
 use criterion::{criterion_group, criterion_main, BenchmarkId, Criterion, Throughput};
 use ezkl::circuit::modules::poseidon::spec::{PoseidonSpec, POSEIDON_RATE, POSEIDON_WIDTH};
 use ezkl::circuit::modules::poseidon::{PoseidonChip, PoseidonConfig};
 use ezkl::circuit::modules::Module;
 use ezkl::circuit::*;
 use ezkl::pfsys::create_keys;
-use ezkl::pfsys::create_proof_circuit_kzg;
+use ezkl::pfsys::create_proof_circuit;
 use ezkl::pfsys::srs::gen_srs;
 use ezkl::pfsys::TranscriptType;
 use ezkl::tensor::*;
 use halo2_proofs::circuit::Value;
 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::{
     arithmetic::Field,
@@ -18,6 +21,7 @@ use halo2_proofs::{
 };
 use halo2curves::bn256::{Bn256, Fr};
 use rand::rngs::OsRng;
+use snark_verifier::system::halo2::transcript::evm::EvmTranscript;
 
 const L: usize = 10;
 
@@ -46,7 +50,7 @@ impl Circuit<Fr> for MyCircuit {
     ) -> Result<(), Error> {
         let chip: PoseidonChip<PoseidonSpec, POSEIDON_WIDTH, POSEIDON_RATE, L> =
             PoseidonChip::new(config);
-        chip.layout(&mut layouter, &[self.image.clone()], 0)?;
+        chip.layout(&mut layouter, &[self.image.clone()], 0, &mut HashMap::new())?;
         Ok(())
     }
 }
@@ -62,7 +66,7 @@ fn runposeidon(c: &mut Criterion) {
         let params = gen_srs::<KZGCommitmentScheme<_>>(k);
 
         let message = (0..*size).map(|_| Fr::random(OsRng)).collect::<Vec<_>>();
-        let output =
+        let _output =
             PoseidonChip::<PoseidonSpec, POSEIDON_WIDTH, POSEIDON_RATE, L>::run(message.to_vec())
                 .unwrap();
 
@@ -76,25 +80,35 @@ fn runposeidon(c: &mut Criterion) {
         group.throughput(Throughput::Elements(*size as u64));
         group.bench_with_input(BenchmarkId::new("pk", size), &size, |b, &_| {
             b.iter(|| {
-                create_keys::<KZGCommitmentScheme<Bn256>, Fr, MyCircuit>(&circuit, &params, true)
+                create_keys::<KZGCommitmentScheme<Bn256>, MyCircuit>(&circuit, &params, true)
                     .unwrap();
             });
         });
 
-        let pk = create_keys::<KZGCommitmentScheme<Bn256>, Fr, MyCircuit>(&circuit, &params, true)
-            .unwrap();
+        let pk =
+            create_keys::<KZGCommitmentScheme<Bn256>, MyCircuit>(&circuit, &params, true).unwrap();
 
         group.throughput(Throughput::Elements(*size as u64));
         group.bench_with_input(BenchmarkId::new("prove", size), &size, |b, &_| {
             b.iter(|| {
-                let prover = create_proof_circuit_kzg(
+                let prover = create_proof_circuit::<
+                    KZGCommitmentScheme<_>,
+                    MyCircuit,
+                    ProverSHPLONK<_>,
+                    VerifierSHPLONK<_>,
+                    SingleStrategy<_>,
+                    _,
+                    EvmTranscript<_, _, _, _>,
+                    EvmTranscript<_, _, _, _>,
+                >(
                     circuit.clone(),
+                    vec![],
                     &params,
-                    Some(output[0].clone()),
                     &pk,
-                    TranscriptType::EVM,
-                    SingleStrategy::new(&params),
                     CheckMode::UNSAFE,
+                    ezkl::Commitments::KZG,
+                    TranscriptType::EVM,
+                    None,
                     None,
                 );
                 prover.unwrap();

benches/relu.rs

@@ -2,11 +2,12 @@ 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::pfsys::create_proof_circuit_kzg;
+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},
@@ -14,6 +15,7 @@ use halo2_proofs::{
 };
 use halo2curves::bn256::{Bn256, Fr};
 use rand::Rng;
+use snark_verifier::system::halo2::transcript::evm::EvmTranscript;
 
 const BITS: Range = (-32768, 32768);
 static mut LEN: usize = 4;
@@ -91,25 +93,35 @@ fn runrelu(c: &mut Criterion) {
         group.throughput(Throughput::Elements(len as u64));
         group.bench_with_input(BenchmarkId::new("pk", len), &len, |b, &_| {
             b.iter(|| {
-                create_keys::<KZGCommitmentScheme<Bn256>, Fr, NLCircuit>(&circuit, &params, true)
+                create_keys::<KZGCommitmentScheme<Bn256>, NLCircuit>(&circuit, &params, true)
                     .unwrap();
             });
         });
 
-        let pk = create_keys::<KZGCommitmentScheme<Bn256>, Fr, 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_kzg(
+                let prover = create_proof_circuit::<
+                    KZGCommitmentScheme<_>,
+                    NLCircuit,
+                    ProverSHPLONK<_>,
+                    VerifierSHPLONK<_>,
+                    SingleStrategy<_>,
+                    _,
+                    EvmTranscript<_, _, _, _>,
+                    EvmTranscript<_, _, _, _>,
+                >(
                     circuit.clone(),
+                    vec![],
                     &params,
-                    None,
                     &pk,
+                    CheckMode::UNSAFE,
+                    ezkl::Commitments::KZG,
                     TranscriptType::EVM,
-                    SingleStrategy::new(&params),
-                    CheckMode::SAFE,
+                    None,
                     None,
                 );
                 prover.unwrap();

docs/python/build.sh

@@ -0,0 +1,2 @@
+#!/bin/sh
+sphinx-build ./src build

docs/python/requirements-docs.txt

@@ -0,0 +1,4 @@
+ezkl==10.3.1
+sphinx
+sphinx-rtd-theme
+sphinxcontrib-napoleon

docs/python/src/conf.py

@@ -0,0 +1,29 @@
+import ezkl
+
+project = 'ezkl'
+release = '10.3.1'
+version = release
+
+
+extensions = [
+    'sphinx.ext.autodoc',
+    'sphinx.ext.autosummary',
+    'sphinx.ext.intersphinx',
+    'sphinx.ext.todo',
+    'sphinx.ext.inheritance_diagram',
+    'sphinx.ext.autosectionlabel',
+    'sphinx.ext.napoleon',
+    'sphinx_rtd_theme',
+]
+
+autosummary_generate = True
+autosummary_imported_members = True
+
+templates_path = ['_templates']
+exclude_patterns = []
+
+# -- Options for HTML output -------------------------------------------------
+# https://www.sphinx-doc.org/en/master/usage/configuration.html#options-for-html-output
+
+html_theme = 'sphinx_rtd_theme'
+html_static_path = ['_static']

docs/python/src/index.rst

@@ -0,0 +1,11 @@
+.. extension documentation master file, created by
+   sphinx-quickstart on Mon Jun 19 15:02:05 2023.
+   You can adapt this file completely to your liking, but it should at least
+   contain the root `toctree` directive.
+
+ezkl python bindings
+================================================
+
+.. automodule:: ezkl
+   :members:
+   :undoc-members:

examples/conv2d_mnist/main.rs

@@ -203,8 +203,8 @@ where
                     let mut region = RegionCtx::new(region, 0, NUM_INNER_COLS);
 
                     let op = PolyOp::Conv {
-                        padding: [(PADDING, PADDING); 2],
-                        stride: (STRIDE, STRIDE),
+                        padding: vec![(PADDING, PADDING); 2],
+                        stride: vec![STRIDE; 2],
                     };
                     let x = config
                         .layer_config

examples/notebooks/ezkl_demo.ipynb

@@ -696,10 +696,12 @@
         "for i, value in enumerate(proof[\"instances\"]):\n",
         "    for j, field_element in enumerate(value):\n",
         "        onchain_input_array.append(ezkl.felt_to_big_endian(field_element))\n",
-        "        formatted_output += str(onchain_input_array[-1])\n",
+        "        formatted_output += '\"' + str(onchain_input_array[-1]) + '\"'\n",
         "        if j != len(value) - 1:\n",
         "            formatted_output += \", \"\n",
-        "    formatted_output += \"]\"\n",
+        "    if i != len(proof[\"instances\"]) - 1:\n",
+        "        formatted_output += \", \"\n",
+        "formatted_output += \"]\"\n",
         "\n",
         "# This will be the values you use onchain\n",
         "# copy them over to remix and see if they verify\n",

examples/notebooks/generalized_inverse.ipynb

@@ -10,7 +10,7 @@
         "\n",
         "## Generalized Inverse\n",
         "\n",
-        "We show how to use EZKL to prove that we know matrices $A$ and its generalized inverse $B$. Since these are large we deal with the KZG commitments, with $a$ the kzgcommit of $A$, $b$ the kzgcommit of $B$, and $ABA = A$.\n"
+        "We show how to use EZKL to prove that we know matrices $A$ and its generalized inverse $B$. Since these are large we deal with the KZG commitments, with $a$ the polycommit of $A$, $b$ the polycommit of $B$, and $ABA = A$.\n"
       ]
     },
     {
@@ -77,7 +77,7 @@
       "outputs": [],
       "source": [
         "gip_run_args = ezkl.PyRunArgs()\n",
-        "gip_run_args.input_visibility = \"kzgcommit\"  # matrix and generalized inverse commitments\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)"
       ]
@@ -340,4 +340,4 @@
   },
   "nbformat": 4,
   "nbformat_minor": 5
-}
+}

examples/notebooks/hashed_vis.ipynb

@@ -161,7 +161,7 @@
                 "- `fixed`: known to the prover and verifier (as a commit), but not modifiable by the prover.\n",
                 "- `hashed`: the hash pre-image is known to the prover, the prover and verifier know the hash. The prover proves that the they know the pre-image to the hash. \n",
                 "- `encrypted`: the non-encrypted element and the secret key used for decryption are known to the prover. The prover and the verifier know the encrypted element, the public key used to encrypt, and the hash of the decryption hey. The prover proves that they know the pre-image of the hashed decryption key and that this key can in fact decrypt the encrypted message.\n",
-                "- `kzgcommit`: unblinded advice column which generates a kzg commitment. This doesn't appear in the instances of the circuit and must instead be inserted directly within the proof bytes.  \n",
+                "- `polycommit`: unblinded advice column which generates a kzg commitment. This doesn't appear in the instances of the circuit and must instead be inserted directly within the proof bytes.  \n",
                 "\n",
                 "\n",
                 "Here we create the following setup:\n",
@@ -510,4 +510,4 @@
     },
     "nbformat": 4,
     "nbformat_minor": 2
-}
+}

examples/notebooks/keras_simple_demo.ipynb

@@ -67,6 +67,7 @@
                 "model.add(Dense(128, activation='relu'))\n",
                 "model.add(Dropout(0.5))\n",
                 "model.add(Dense(10, activation='softmax'))\n",
+                "model.output_names=['output']\n",
                 "\n",
                 "\n",
                 "# Train the model as you like here (skipped for brevity)\n",

examples/notebooks/kzg_vis.ipynb

@@ -154,11 +154,11 @@
                 "- `fixed`: known to the prover and verifier (as a commit), but not modifiable by the prover.\n",
                 "- `hashed`: the hash pre-image is known to the prover, the prover and verifier know the hash. The prover proves that the they know the pre-image to the hash. \n",
                 "- `encrypted`: the non-encrypted element and the secret key used for decryption are known to the prover. The prover and the verifier know the encrypted element, the public key used to encrypt, and the hash of the decryption hey. The prover proves that they know the pre-image of the hashed decryption key and that this key can in fact decrypt the encrypted message.\n",
-                "- `kzgcommit`: unblinded advice column which generates a kzg commitment. This doesn't appear in the instances of the circuit and must instead be modified directly within the proof bytes.  \n",
+                "- `polycommit`: unblinded advice column which generates a kzg commitment. This doesn't appear in the instances of the circuit and must instead be modified directly within the proof bytes.  \n",
                 "\n",
                 "Here we create the following setup:\n",
-                "- `input_visibility`: \"kzgcommit\"\n",
-                "- `param_visibility`: \"kzgcommit\"\n",
+                "- `input_visibility`: \"polycommit\"\n",
+                "- `param_visibility`: \"polycommit\"\n",
                 "- `output_visibility`: public\n",
                 "\n",
                 "We encourage you to play around with other setups :) \n",
@@ -186,8 +186,8 @@
                 "data_path = os.path.join('input.json')\n",
                 "\n",
                 "run_args = ezkl.PyRunArgs()\n",
-                "run_args.input_visibility = \"kzgcommit\"\n",
-                "run_args.param_visibility = \"kzgcommit\"\n",
+                "run_args.input_visibility = \"polycommit\"\n",
+                "run_args.param_visibility = \"polycommit\"\n",
                 "run_args.output_visibility = \"public\"\n",
                 "run_args.variables = [(\"batch_size\", 1)]\n",
                 "\n",
@@ -512,4 +512,4 @@
     },
     "nbformat": 4,
     "nbformat_minor": 2
-}
+}

examples/notebooks/mnist_gan.ipynb

@@ -38,7 +38,7 @@
                 "import logging\n",
                 "\n",
                 "import tensorflow as tf\n",
-                "from tensorflow.keras.optimizers.legacy import Adam\n",
+                "from tensorflow.keras.optimizers import Adam\n",
                 "from tensorflow.keras.layers import *\n",
                 "from tensorflow.keras.models import Model\n",
                 "from tensorflow.keras.datasets import mnist\n",
@@ -71,9 +71,11 @@
             },
             "outputs": [],
             "source": [
-                "opt = Adam()\n",
                 "ZDIM = 100\n",
                 "\n",
+                "opt = Adam()\n",
+                "\n",
+                "\n",
                 "# discriminator\n",
                 "# 0 if it's fake, 1 if it's real\n",
                 "x = in1 = Input((28,28))\n",
@@ -114,8 +116,11 @@
                 "\n",
                 "gm = Model(in1, x)\n",
                 "gm.compile('adam', 'mse')\n",
+                "gm.output_names=['output']\n",
                 "gm.summary()\n",
                 "\n",
+                "opt = Adam()\n",
+                "\n",
                 "# GAN\n",
                 "dm.trainable = False\n",
                 "x = dm(gm.output)\n",
@@ -415,7 +420,7 @@
             "name": "python",
             "nbconvert_exporter": "python",
             "pygments_lexer": "ipython3",
-            "version": "3.9.15"
+            "version": "3.12.2"
         }
     },
     "nbformat": 4,

examples/notebooks/mnist_vae.ipynb

@@ -349,6 +349,8 @@
                 "z_log_var = Dense(ZDIM)(x)\n",
                 "z = Lambda(lambda x: x[0] + K.exp(0.5 * x[1]) * K.random_normal(shape=K.shape(x[0])))([z_mu, z_log_var])\n",
                 "dec = get_decoder()\n",
+                "dec.output_names=['output']\n",
+                "\n",
                 "out = dec(z)\n",
                 "\n",
                 "mse_loss = mse(Reshape((28*28,))(in1), Reshape((28*28,))(out)) * 28 * 28\n",

examples/notebooks/proof_splitting.ipynb

@@ -208,7 +208,7 @@
                 "- `private`: known only to the prover\n",
                 "- `hashed`: the hash pre-image is known to the prover, the prover and verifier know the hash. The prover proves that the they know the pre-image to the hash. \n",
                 "- `encrypted`: the non-encrypted element and the secret key used for decryption are known to the prover. The prover and the verifier know the encrypted element, the public key used to encrypt, and the hash of the decryption hey. The prover proves that they know the pre-image of the hashed decryption key and that this key can in fact decrypt the encrypted message.\n",
-                "- `kzgcommit`: unblinded advice column which generates a kzg commitment. This doesn't appear in the instances of the circuit and must instead be modified directly within the proof bytes.  \n",
+                "- `polycommit`: unblinded advice column which generates a kzg commitment. This doesn't appear in the instances of the circuit and must instead be modified directly within the proof bytes.  \n",
                 "\n",
                 "Here we create the following setup:\n",
                 "- `input_visibility`: \"public\"\n",
@@ -234,7 +234,7 @@
                 "run_args.input_scale = 2\n",
                 "run_args.logrows = 8\n",
                 "\n",
-                "ezkl.get_srs(logrows=run_args.logrows)"
+                "ezkl.get_srs(logrows=run_args.logrows, commitment=ezkl.PyCommitments.KZG)"
             ]
         },
         {
@@ -385,9 +385,9 @@
                 "###  KZG commitment intermediate calculations\n",
                 "\n",
                 "This time the visibility parameters are:\n",
-                "- `input_visibility`: \"kzgcommit\"\n",
+                "- `input_visibility`: \"polycommit\"\n",
                 "- `param_visibility`: \"public\"\n",
-                "- `output_visibility`: kzgcommit"
+                "- `output_visibility`: polycommit"
             ]
         },
         {
@@ -399,9 +399,9 @@
                 "import ezkl\n",
                 "\n",
                 "run_args = ezkl.PyRunArgs()\n",
-                "run_args.input_visibility = \"kzgcommit\"\n",
+                "run_args.input_visibility = \"polycommit\"\n",
                 "run_args.param_visibility = \"fixed\"\n",
-                "run_args.output_visibility = \"kzgcommit\"\n",
+                "run_args.output_visibility = \"polycommit\"\n",
                 "run_args.variables = [(\"batch_size\", 1)]\n",
                 "run_args.input_scale = 2\n",
                 "run_args.logrows = 8\n"

examples/notebooks/random_forest.ipynb

@@ -61,11 +61,10 @@
                 "from sklearn.datasets import load_iris\n",
                 "from sklearn.model_selection import train_test_split\n",
                 "from sklearn.ensemble import RandomForestClassifier as Rf\n",
-                "import sk2torch\n",
                 "import torch\n",
                 "import ezkl\n",
                 "import os\n",
-                "from torch import nn\n",
+                "from hummingbird.ml import convert\n",
                 "\n",
                 "\n",
                 "\n",
@@ -77,28 +76,12 @@
                 "clr.fit(X_train, y_train)\n",
                 "\n",
                 "\n",
-                "trees = []\n",
-                "for tree in clr.estimators_:\n",
-                "    trees.append(sk2torch.wrap(tree))\n",
                 "\n",
-                "\n",
-                "class RandomForest(nn.Module):\n",
-                "    def __init__(self, trees):\n",
-                "        super(RandomForest, self).__init__()\n",
-                "        self.trees = nn.ModuleList(trees)\n",
-                "\n",
-                "    def forward(self, x):\n",
-                "        out = self.trees[0](x)\n",
-                "        for tree in self.trees[1:]:\n",
-                "            out += tree(x)\n",
-                "        return out / len(self.trees)\n",
-                "\n",
-                "\n",
-                "torch_rf = RandomForest(trees)\n",
+                "torch_rf = convert(clr, 'torch')\n",
                 "# assert predictions from torch are = to sklearn \n",
                 "diffs = []\n",
                 "for i in range(len(X_test)):\n",
-                "    torch_pred = torch_rf(torch.tensor(X_test[i].reshape(1, -1)))\n",
+                "    torch_pred = torch_rf.predict(torch.tensor(X_test[i].reshape(1, -1)))\n",
                 "    sk_pred = clr.predict(X_test[i].reshape(1, -1))\n",
                 "    diffs.append(torch_pred[0].round() - sk_pred[0])\n",
                 "\n",
@@ -134,14 +117,12 @@
                 "\n",
                 "# export to onnx format\n",
                 "\n",
-                "torch_rf.eval()\n",
-                "\n",
                 "# Input to the model\n",
                 "shape = X_train.shape[1:]\n",
                 "x = torch.rand(1, *shape, requires_grad=False)\n",
-                "torch_out = torch_rf(x)\n",
+                "torch_out = torch_rf.predict(x)\n",
                 "# Export the model\n",
-                "torch.onnx.export(torch_rf,               # model being run\n",
+                "torch.onnx.export(torch_rf.model,               # 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",
@@ -158,7 +139,7 @@
                 "\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",
+                "            output_data=[o.reshape([-1]).tolist() for o in torch_out])\n",
                 "\n",
                 "# Serialize data into file:\n",
                 "json.dump(data, open(\"input.json\", 'w'))\n"
@@ -321,7 +302,7 @@
             "name": "python",
             "nbconvert_exporter": "python",
             "pygments_lexer": "ipython3",
-            "version": "3.9.15"
+            "version": "3.12.2"
         }
     },
     "nbformat": 4,

examples/notebooks/simple_demo_aggregated_proofs.ipynb

@@ -275,7 +275,6 @@
                 "        proof_path,\n",
                 "        settings_path,\n",
                 "        vk_path,\n",
-                "        \n",
                 "    )\n",
                 "\n",
                 "assert res == True\n",
@@ -291,7 +290,7 @@
             "source": [
                 "# Generate a larger SRS. This is needed for the aggregated proof\n",
                 "\n",
-                "res = ezkl.get_srs(settings_path=None, logrows=21)"
+                "res = ezkl.get_srs(settings_path=None, logrows=21, commitment=ezkl.PyCommitments.KZG)"
             ]
         },
         {

examples/notebooks/solvency.ipynb

@@ -9,7 +9,7 @@
       "source": [
         "## Solvency demo\n",
         "\n",
-        "Here we create a demo of a solvency calculation in the manner of [summa-solvency](https://github.com/summa-dev/summa-solvency). The aim here is to demonstrate the use of the new kzgcommit method detailed [here](https://blog.ezkl.xyz/post/commits/). \n",
+        "Here we create a demo of a solvency calculation in the manner of [summa-solvency](https://github.com/summa-dev/summa-solvency). The aim here is to demonstrate the use of the new polycommit method detailed [here](https://blog.ezkl.xyz/post/commits/). \n",
         "\n",
         "In this setup:\n",
         "- the commitments to users, respective balances, and total balance are known are publicly known to the prover and verifier. \n",
@@ -177,10 +177,10 @@
         "- `private`: known only to the prover\n",
         "- `hashed`: the hash pre-image is known to the prover, the prover and verifier know the hash. The prover proves that the they know the pre-image to the hash. \n",
         "- `encrypted`: the non-encrypted element and the secret key used for decryption are known to the prover. The prover and the verifier know the encrypted element, the public key used to encrypt, and the hash of the decryption hey. The prover proves that they know the pre-image of the hashed decryption key and that this key can in fact decrypt the encrypted message.\n",
-        "- `kzgcommit`: unblinded advice column which generates a kzg commitment. This doesn't appear in the instances of the circuit and must instead be modified directly within the proof bytes.  \n",
+        "- `polycommit`: unblinded advice column which generates a kzg commitment. This doesn't appear in the instances of the circuit and must instead be modified directly within the proof bytes.  \n",
         "\n",
         "Here we create the following setup:\n",
-        "- `input_visibility`: \"kzgcommit\"\n",
+        "- `input_visibility`: \"polycommit\"\n",
         "- `param_visibility`: \"public\"\n",
         "- `output_visibility`: public\n",
         "\n",
@@ -202,8 +202,8 @@
       "outputs": [],
       "source": [
         "run_args = ezkl.PyRunArgs()\n",
-        "# \"kzgcommit\" 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 = \"kzgcommit\"\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",
         "# the parameters are public\n",
         "run_args.param_visibility = \"fixed\"\n",
         "# the output is public (this is the inequality test)\n",
@@ -515,4 +515,4 @@
   },
   "nbformat": 4,
   "nbformat_minor": 5
-}
+}

examples/notebooks/tictactoe_autoencoder.ipynb

@@ -13,7 +13,7 @@
         },
         {
             "cell_type": "code",
-            "execution_count": null,
+            "execution_count": 1,
             "metadata": {},
             "outputs": [],
             "source": [
@@ -57,7 +57,7 @@
         },
         {
             "cell_type": "code",
-            "execution_count": null,
+            "execution_count": 2,
             "metadata": {},
             "outputs": [],
             "source": [
@@ -119,7 +119,7 @@
         },
         {
             "cell_type": "code",
-            "execution_count": null,
+            "execution_count": 3,
             "metadata": {},
             "outputs": [],
             "source": [
@@ -163,7 +163,7 @@
         },
         {
             "cell_type": "code",
-            "execution_count": null,
+            "execution_count": 4,
             "metadata": {},
             "outputs": [],
             "source": [
@@ -217,7 +217,7 @@
         },
         {
             "cell_type": "code",
-            "execution_count": null,
+            "execution_count": 6,
             "metadata": {},
             "outputs": [],
             "source": [
@@ -227,6 +227,10 @@
                 "        self.length = self.compute_length(self.file_good)\n",
                 "        self.data = self.load_data(self.file_good)\n",
                 "\n",
+                "    def __iter__(self):\n",
+                "        for i in range(len(self.data)):\n",
+                "            yield self.data[i]\n",
+                "\n",
                 "    def parse_json_object(self, line):\n",
                 "        try:\n",
                 "            return json.loads(line)\n",
@@ -749,7 +753,7 @@
             "name": "python",
             "nbconvert_exporter": "python",
             "pygments_lexer": "ipython3",
-            "version": "3.9.15"
+            "version": "3.12.2"
         }
     },
     "nbformat": 4,

examples/notebooks/tictactoe_binary_classification.ipynb

@@ -209,6 +209,11 @@
                 "        self.length = self.compute_length(self.file_good, self.file_bad)\n",
                 "        self.data = self.load_data(self.file_good, self.file_bad)\n",
                 "\n",
+                "    def __iter__(self):\n",
+                "        for i in range(len(self.data)):\n",
+                "            yield self.data[i]\n",
+                "\n",
+                "\n",
                 "    def parse_json_object(self, line):\n",
                 "        try:\n",
                 "            return json.loads(line)\n",
@@ -637,7 +642,7 @@
             "name": "python",
             "nbconvert_exporter": "python",
             "pygments_lexer": "ipython3",
-            "version": "3.9.15"
+            "version": "3.12.2"
         }
     },
     "nbformat": 4,

examples/onnx/1l_lppool/gen.py

@@ -0,0 +1,40 @@
+from torch import nn
+import torch
+import json
+
+class Model(nn.Module):
+    def __init__(self):
+        super(Model, self).__init__()
+        self.layer = nn.LPPool2d(2, 1, (1, 1))
+
+    def forward(self, x):
+        return self.layer(x)[0]
+
+
+circuit = Model()
+
+x = torch.empty(1, 3, 2, 2).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/1l_lppool/input.json

@@ -0,0 +1 @@
+{"input_data": [[0.7549541592597961, 0.990360677242279, 0.9473411440849304, 0.3951031565666199, 0.8500555753707886, 0.9352139830589294, 0.11867779493331909, 0.9493132829666138, 0.6588345766067505, 0.1933223009109497, 0.12139874696731567, 0.8547163605690002]]}

examples/onnx/celu/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 = nn.CELU()(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/celu/input.json

@@ -0,0 +1 @@
+{"input_data": [[0.35387128591537476, 0.030473172664642334, 0.08707714080810547, 0.2429301142692566, 0.45228832960128784, 0.496021032333374, 0.13245105743408203, 0.8497090339660645]]}

examples/onnx/clip/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 = torch.clamp(x, min=0.4, max=0.8)
+        return m
+
+
+circuit = MyModel()
+
+x = torch.empty(1, 2, 2, 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/clip/input.json

@@ -0,0 +1 @@
+{"input_data": [[0.03297048807144165, 0.46362626552581787, 0.6044231057167053, 0.4949902892112732, 0.48823297023773193, 0.6798646450042725, 0.6824942231178284, 0.03491640090942383, 0.19608813524246216, 0.24129939079284668, 0.9769315123558044, 0.6306831240653992, 0.7690497636795044, 0.252221941947937, 0.9167693853378296, 0.3882681131362915, 0.9307044148445129, 0.33559417724609375, 0.7815426588058472, 0.3435332179069519, 0.7871478796005249, 0.12240773439407349, 0.5295405983924866, 0.4874419569969177, 0.08262640237808228, 0.1124718189239502, 0.5834914445877075, 0.30927878618240356, 0.48899340629577637, 0.9376634955406189, 0.21893149614334106, 0.526070773601532]]}

examples/onnx/clip/network.onnx

@@ -0,0 +1,24 @@
+pytorch2.2.1:±
+?/Constant_output_0	/Constant"Constant*
+value*
JÍÌÌ> 

+C/Constant_1_output_0/Constant_1"Constant*
+value*
JÍÌL? 

+F
+input
+/Constant_output_0
+/Constant_1_output_0output/Clip"Clip
+main_graphZ)
+input 
+

+
+batch_size
+
+
+b*
+output 
+

+
+batch_size
+
+
+B

examples/onnx/gru/gen.py

@@ -0,0 +1,41 @@
+import random
+import math
+import numpy as np
+
+import torch
+from torch import nn
+import torch.nn.functional as F
+import json
+
+
+model = nn.GRU(3, 3)  # Input dim is 3, output dim is 3
+x = torch.randn(1, 3)  # make a sequence of length 5
+
+print(x)
+
+# Flips the neural net into inference mode
+model.eval()
+model.to('cpu')
+
+# Export the model
+torch.onnx.export(model,               # model being run
+                  # model input (or a tuple for multiple inputs)
+                  x,
+                  # where to save the model (can be a file or file-like object)
+                  "network.onnx",
+                  export_params=True,        # store the trained parameter weights inside the model file
+                  opset_version=10,          # 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'}})
+
+data_array = ((x).detach().numpy()).reshape([-1]).tolist()
+
+data_json = dict(input_data=[data_array])
+
+print(data_json)
+
+# Serialize data into file:
+json.dump(data_json, open("input.json", 'w'))

examples/onnx/gru/input.json

@@ -0,0 +1 @@
+{"input_data": [[0.4145222008228302, -0.4043896496295929, 0.7545749545097351]]}

examples/onnx/hard_max/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.argmax(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/hard_max/input.json

@@ -0,0 +1 @@
+{"input_data": [[0.5505883693695068, 0.0766521692276001, 0.12006187438964844, 0.9497959017753601, 0.9100563526153564, 0.968717098236084, 0.5978299379348755, 0.9419963359832764]]}

examples/onnx/hard_sigmoid/gen.py

@@ -9,7 +9,7 @@ class MyModel(nn.Module):
         super(MyModel, self).__init__()
 
     def forward(self, x):          
-        m = nn.Logsoftmax()(x)
+        m = nn.Hardsigmoid()(x)
         
         return m 
 

examples/onnx/hard_sigmoid/input.json

@@ -1 +1 @@
-{"input_data": [[0.2971532940864563, 0.3465197682380676, 0.05381882190704346, 0.058654189109802246, 0.014198064804077148, 0.06088751554489136, 0.1723427176475525, 0.5115123987197876]]}
+{"input_data": [[0.8326942324638367, 0.2796096205711365, 0.600328266620636, 0.3701696991920471, 0.17832040786743164, 0.6247223019599915, 0.501872718334198, 0.6961578726768494]]}

examples/onnx/hard_sigmoid/network.onnx

@@ -1,4 +1,4 @@
-pytorch2.1.0:<3A>

+pytorch2.2.1:<3A>

 ;
 inputoutput/HardSigmoid"HardSigmoid*
 alpha«ª*> 



examples/onnx/hard_swish/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 = nn.Hardswish()(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/hard_swish/input.json

@@ -0,0 +1 @@
+{"input_data": [[0.6996762752532959, 0.42992985248565674, 0.5102168321609497, 0.5540387630462646, 0.8489438891410828, 0.8533616065979004, 0.36736780405044556, 0.5859147310256958]]}

examples/onnx/hard_swish/network.onnx

@@ -0,0 +1,15 @@
+pytorch2.2.1:{
+&
+inputoutput
+/HardSwish"	HardSwish
+main_graphZ!
+input
+

+
+batch_size
+b"
+output
+

+
+batch_size
+B

examples/onnx/log_softmax/gen.py

@@ -9,7 +9,7 @@ class MyModel(nn.Module):
         super(MyModel, self).__init__()
 
     def forward(self, x):          
-        m = nn.Hardsigmoid()(x)
+        m = nn.LogSoftmax()(x)
         
         return m 
 

examples/onnx/logsumexp/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.logsumexp(x, dim=1)
+        
+        return m 
+
+
+circuit = MyModel()
+
+x = torch.empty(1, 2, 2, 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/logsumexp/input.json

@@ -0,0 +1 @@
+{"input_data": [[0.7973018884658813, 0.5245689153671265, 0.34149593114852905, 0.1455438733100891, 0.9482707381248474, 0.4221445322036743, 0.001363217830657959, 0.8736765384674072, 0.42954301834106445, 0.7199509739875793, 0.37641745805740356, 0.5920265316963196, 0.42270803451538086, 0.41761744022369385, 0.603948712348938, 0.7250819802284241, 0.047173500061035156, 0.5115441679954529, 0.3743387460708618, 0.16794061660766602, 0.5352339148521423, 0.037976861000061035, 0.65323406457901, 0.5585184097290039, 0.10559147596359253, 0.07827490568161011, 0.6717077493667603, 0.6480781435966492, 0.9780838489532471, 0.8353415131568909, 0.6491701006889343, 0.6573048233985901]]}

examples/onnx/lstm_large/input.json

@@ -0,0 +1,13 @@
+{
+  "input_data": [
+    [
+      0.8894134163856506,
+      0.8894201517105103
+    ]
+  ],
+  "output_data": [
+    [
+      0.8436377
+    ]
+  ]
+}

examples/onnx/mish/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 = nn.Mish()(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/mish/input.json

@@ -0,0 +1 @@
+{"input_data": [[0.18563222885131836, 0.4843214750289917, 0.9991059899330139, 0.02534431219100952, 0.8105666041374207, 0.9658406376838684, 0.681107759475708, 0.5365872979164124]]}

examples/onnx/mish/network.onnx

@@ -0,0 +1,19 @@
+pytorch2.2.1:ä

+0
+input/Softplus_output_0	/Softplus"Softplus
+1
+/Softplus_output_0/Tanh_output_0/Tanh"Tanh
+*
+input
+/Tanh_output_0output/Mul"Mul
+main_graphZ!
+input
+

+
+batch_size
+b"
+output
+

+
+batch_size
+B

examples/onnx/reducel1/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.norm(x, p=1, dim=1)
+        
+        return m 
+
+
+circuit = MyModel()
+
+x = torch.empty(1, 2, 2, 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/reducel1/input.json

@@ -0,0 +1 @@
+{"input_data": [[0.02284395694732666, 0.7941043376922607, 0.07971876859664917, 0.8898420929908752, 0.8233054280281067, 0.11066079139709473, 0.4424799084663391, 0.4355071783065796, 0.6723723411560059, 0.6818525195121765, 0.8726171851158142, 0.17742449045181274, 0.054257750511169434, 0.5775953531265259, 0.7758923172950745, 0.8431423306465149, 0.7602444887161255, 0.29686522483825684, 0.22489851713180542, 0.0675363540649414, 0.981339693069458, 0.15771394968032837, 0.5801441669464111, 0.9044001698493958, 0.49266451597213745, 0.42621421813964844, 0.35345613956451416, 0.042848050594329834, 0.6908614039421082, 0.5422852039337158, 0.01975083351135254, 0.5772860050201416]]}

examples/onnx/reducel2/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.norm(x, p=2, dim=1)
+        
+        return m 
+
+
+circuit = MyModel()
+
+x = torch.empty(1, 2, 2, 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/reducel2/input.json

@@ -0,0 +1 @@
+{"input_data": [[0.8709188103675842, 0.11553549766540527, 0.27376580238342285, 0.7518517971038818, 0.7879393100738525, 0.8765475749969482, 0.14315760135650635, 0.8982420563697815, 0.7274006605148315, 0.39007169008255005, 0.729040801525116, 0.11306107044219971, 0.658822774887085, 0.666404664516449, 0.3001367449760437, 0.45343858003616333, 0.7460223436355591, 0.7423691749572754, 0.7544230818748474, 0.5674425959587097, 0.8728761672973633, 0.27062875032424927, 0.1595977544784546, 0.22975260019302368, 0.6711723208427429, 0.8265992403030396, 0.48679041862487793, 0.689740777015686, 0.330846905708313, 0.5630669593811035, 0.8058932423591614, 0.5802426338195801]]}

examples/onnx/tril/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.triu(x)
+        
+        return m 
+
+
+circuit = MyModel()
+
+x = torch.empty(1, 3, 3).uniform_(0, 5)
+
+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/tril/input.json

@@ -0,0 +1 @@
+{"input_data": [[0.4870188236236572, 2.275230646133423, 3.126268148422241, 0.6412187218666077, 0.9967470169067383, 1.9814395904541016, 1.6355383396148682, 0.6397527456283569, 0.7825168967247009]]}

examples/onnx/triu/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.tril(x)
+        
+        return m 
+
+
+circuit = MyModel()
+
+x = torch.empty(1, 3, 3).uniform_(0, 5)
+
+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/triu/input.json

@@ -0,0 +1 @@
+{"input_data": [[0.2898547053337097, 1.8070811033248901, 0.30266255140304565, 3.00581955909729, 0.5379888415336609, 1.7057424783706665, 2.415961265563965, 0.589233934879303, 0.03824889659881592]]}

in-browser-evm-verifier/package.json

@@ -17,7 +17,7 @@
     "clean": "rm -r dist || true",
     "build:commonjs": "tsc --project tsconfig.commonjs.json && resolve-tspaths -p tsconfig.commonjs.json",
     "build:esm": "tsc --project tsconfig.esm.json && resolve-tspaths -p tsconfig.esm.json",
-    "build": "pnpm run clean && pnpm run build:commonjs && pnpm run build:esm"
+    "build": "npm run clean && npm run build:commonjs && npm run build:esm"
   },
   "dependencies": {
     "@ethereumjs/common": "^4.0.0",

install_ezkl_cli.sh

@@ -36,7 +36,7 @@ if  [ "$(which ezkl)s" != "s" ] && [ "$(which ezkl)" != "$EZKL_DIR/ezkl" ] ; the
     exit 1
 fi
 
-if [[ ":$PATH:" != *":${EZKl_DIR}:"* ]]; then
+if [[ ":$PATH:" != *":${EZKL_DIR}:"* ]]; then
     # Add the ezkl directory to the path and ensure the old PATH variables remain.
     echo >> $PROFILE && echo "export PATH=\"\$PATH:$EZKL_DIR\"" >> $PROFILE
 fi

pyproject.toml

@@ -1,5 +1,5 @@
 [build-system]
-requires = ["maturin>=0.14,<0.15"]
+requires = ["maturin>=1.0,<2.0"]
 build-backend = "maturin"
 
 [tool.pytest.ini_options]

requirements.txt

@@ -1,14 +1,14 @@
-attrs==22.2.0
-exceptiongroup==1.1.1
-importlib-metadata==6.1.0
+attrs==23.2.0
+exceptiongroup==1.2.0
+importlib-metadata==7.1.0
 iniconfig==2.0.0
-maturin==1.0.1
-packaging==23.0
-pluggy==1.0.0
-pytest==7.2.2
+maturin==1.5.1
+packaging==24.0
+pluggy==1.4.0
+pytest==8.1.1
 tomli==2.0.1
-typing-extensions==4.5.0
-zipp==3.15.0
-onnx==1.14.1
-onnxruntime==1.14.1
-numpy==1.21.6
+typing-extensions==4.10.0
+zipp==3.18.1
+onnx==1.15.0
+onnxruntime==1.17.1
+numpy==1.26.4

rust-toolchain

@@ -1,3 +1,3 @@
 [toolchain]
-channel = "nightly-2023-08-24"
+channel = "nightly-2024-02-06"
 components = ["rustfmt", "clippy"]

src/circuit/modules/mod.rs

@@ -2,7 +2,7 @@
 pub mod poseidon;
 
 ///
-pub mod kzg;
+pub mod polycommit;
 
 ///
 pub mod planner;
@@ -15,6 +15,8 @@ pub use planner::*;
 
 use crate::tensor::{TensorType, ValTensor};
 
+use super::region::ConstantsMap;
+
 /// Module trait used to extend ezkl functionality
 pub trait Module<F: PrimeField + TensorType + PartialOrd> {
     /// Config
@@ -39,6 +41,7 @@ pub trait Module<F: PrimeField + TensorType + PartialOrd> {
         &self,
         layouter: &mut impl Layouter<F>,
         input: &[ValTensor<F>],
+        constants: &mut ConstantsMap<F>,
     ) -> Result<Self::InputAssignments, Error>;
     /// Layout
     fn layout(
@@ -46,6 +49,7 @@ pub trait Module<F: PrimeField + TensorType + PartialOrd> {
         layouter: &mut impl Layouter<F>,
         input: &[ValTensor<F>],
         row_offset: usize,
+        constants: &mut ConstantsMap<F>,
     ) -> Result<ValTensor<F>, Error>;
     /// Number of instance values the module uses every time it is applied
     fn instance_increment_input(&self) -> Vec<usize>;

src/circuit/modules/polycommit.rs

@@ -4,49 +4,49 @@ is already implemented in halo2_gadgets, there is no wrapper chip that makes it
 Thanks to https://github.com/summa-dev/summa-solvency/blob/master/src/chips/poseidon/hash.rs for the inspiration (and also helping us understand how to use this).
 */
 
+use std::collections::HashMap;
+
 // This chip adds a set of advice columns to the gadget Chip to store the inputs of the hash
 use halo2_proofs::halo2curves::bn256::Fr as Fp;
-use halo2_proofs::poly::commitment::{Blind, Params};
-use halo2_proofs::poly::kzg::commitment::ParamsKZG;
+use halo2_proofs::poly::commitment::{Blind, CommitmentScheme, Params};
 use halo2_proofs::{circuit::*, plonk::*};
-use halo2curves::bn256::{Bn256, G1Affine};
+use halo2curves::bn256::G1Affine;
 use halo2curves::group::prime::PrimeCurveAffine;
 use halo2curves::group::Curve;
 use halo2curves::CurveAffine;
 
+use crate::circuit::region::ConstantsMap;
 use crate::tensor::{Tensor, ValTensor, ValType, VarTensor};
 
 use super::Module;
 
-/// The number of instance columns used by the KZG hash function
+/// The number of instance columns used by the PolyCommit hash function
 pub const NUM_INSTANCE_COLUMNS: usize = 0;
-/// The number of advice columns used by the KZG hash function
+/// The number of advice columns used by the PolyCommit hash function
 pub const NUM_INNER_COLS: usize = 1;
 
 #[derive(Debug, Clone)]
-/// WIDTH, RATE and L are const generics for the struct, which represent the width, rate, and number of inputs for the Poseidon hash function, respectively.
-/// This means they are values that are known at compile time and can be used to specialize the implementation of the struct.
-/// The actual chip provided by halo2_gadgets is added to the parent Chip.
-pub struct KZGConfig {
+/// Configuration for the PolyCommit chip
+pub struct PolyCommitConfig {
     ///
-    pub hash_inputs: VarTensor,
+    pub inputs: VarTensor,
 }
 
 type InputAssignments = ();
 
-/// PoseidonChip is a wrapper around the Pow5Chip that adds a set of advice columns to the gadget Chip to store the inputs of the hash
+///
 #[derive(Debug)]
-pub struct KZGChip {
-    config: KZGConfig,
+pub struct PolyCommitChip {
+    config: PolyCommitConfig,
 }
 
-impl KZGChip {
+impl PolyCommitChip {
     /// Commit to the message using the KZG commitment scheme
-    pub fn commit(
-        message: Vec<Fp>,
+    pub fn commit<Scheme: CommitmentScheme<Scalar = Fp, Curve = G1Affine>>(
+        message: Vec<Scheme::Scalar>,
         degree: u32,
         num_unusable_rows: u32,
-        params: &ParamsKZG<Bn256>,
+        params: &Scheme::ParamsProver,
     ) -> Vec<G1Affine> {
         let k = params.k();
         let domain = halo2_proofs::poly::EvaluationDomain::new(degree, k);
@@ -81,14 +81,14 @@ impl KZGChip {
     }
 }
 
-impl Module<Fp> for KZGChip {
-    type Config = KZGConfig;
+impl Module<Fp> for PolyCommitChip {
+    type Config = PolyCommitConfig;
     type InputAssignments = InputAssignments;
     type RunInputs = Vec<Fp>;
     type Params = (usize, usize);
 
     fn name(&self) -> &'static str {
-        "KZG"
+        "PolyCommit"
     }
 
     fn instance_increment_input(&self) -> Vec<usize> {
@@ -102,14 +102,15 @@ impl Module<Fp> for KZGChip {
 
     /// Configuration of the PoseidonChip
     fn configure(meta: &mut ConstraintSystem<Fp>, params: Self::Params) -> Self::Config {
-        let hash_inputs = VarTensor::new_unblinded_advice(meta, params.0, NUM_INNER_COLS, params.1);
-        Self::Config { hash_inputs }
+        let inputs = VarTensor::new_unblinded_advice(meta, params.0, NUM_INNER_COLS, params.1);
+        Self::Config { inputs }
     }
 
     fn layout_inputs(
         &self,
         _: &mut impl Layouter<Fp>,
         _: &[ValTensor<Fp>],
+        _: &mut ConstantsMap<Fp>,
     ) -> Result<Self::InputAssignments, Error> {
         Ok(())
     }
@@ -122,11 +123,24 @@ impl Module<Fp> for KZGChip {
         layouter: &mut impl Layouter<Fp>,
         input: &[ValTensor<Fp>],
         _: usize,
+        constants: &mut ConstantsMap<Fp>,
     ) -> Result<ValTensor<Fp>, Error> {
         assert_eq!(input.len(), 1);
+
+        let local_constants = constants.clone();
         layouter.assign_region(
-            || "kzg commit",
-            |mut region| self.config.hash_inputs.assign(&mut region, 0, &input[0]),
+            || "PolyCommit",
+            |mut region| {
+                let mut local_inner_constants = local_constants.clone();
+                let res = self.config.inputs.assign(
+                    &mut region,
+                    0,
+                    &input[0],
+                    &mut local_inner_constants,
+                )?;
+                *constants = local_inner_constants;
+                Ok(res)
+            },
         )
     }
 
@@ -163,7 +177,7 @@ mod tests {
     }
 
     impl Circuit<Fp> for HashCircuit {
-        type Config = KZGConfig;
+        type Config = PolyCommitConfig;
         type FloorPlanner = ModulePlanner;
         type Params = ();
 
@@ -178,7 +192,7 @@ mod tests {
 
         fn configure(meta: &mut ConstraintSystem<Fp>) -> Self::Config {
             let params = (K, R);
-            KZGChip::configure(meta, params)
+            PolyCommitChip::configure(meta, params)
         }
 
         fn synthesize(
@@ -186,8 +200,13 @@ mod tests {
             config: Self::Config,
             mut layouter: impl Layouter<Fp>,
         ) -> Result<(), Error> {
-            let kzg_chip = KZGChip::new(config);
-            kzg_chip.layout(&mut layouter, &[self.message.clone()], 0);
+            let polycommit_chip = PolyCommitChip::new(config);
+            polycommit_chip.layout(
+                &mut layouter,
+                &[self.message.clone()],
+                0,
+                &mut HashMap::new(),
+            );
 
             Ok(())
         }
@@ -195,7 +214,7 @@ mod tests {
 
     #[test]
     #[ignore]
-    fn kzg_for_a_range_of_input_sizes() {
+    fn polycommit_chip_for_a_range_of_input_sizes() {
         let rng = rand::rngs::OsRng;
 
         #[cfg(not(target_arch = "wasm32"))]
@@ -225,7 +244,7 @@ mod tests {
 
     #[test]
     #[ignore]
-    fn kzg_commit_much_longer_input() {
+    fn polycommit_chip_much_longer_input() {
         #[cfg(not(target_arch = "wasm32"))]
         env_logger::init();
 

src/circuit/modules/poseidon.rs

@@ -18,6 +18,7 @@ use maybe_rayon::slice::ParallelSlice;
 
 use std::marker::PhantomData;
 
+use crate::circuit::region::ConstantsMap;
 use crate::tensor::{Tensor, ValTensor, ValType};
 
 use super::Module;
@@ -172,12 +173,15 @@ impl<S: Spec<Fp, WIDTH, RATE> + Sync, const WIDTH: usize, const RATE: usize, con
         &self,
         layouter: &mut impl Layouter<Fp>,
         message: &[ValTensor<Fp>],
+        constants: &mut ConstantsMap<Fp>,
     ) -> Result<Self::InputAssignments, Error> {
         assert_eq!(message.len(), 1);
         let message = message[0].clone();
 
         let start_time = instant::Instant::now();
 
+        let local_constants = constants.clone();
+
         let res = layouter.assign_region(
             || "load message",
             |mut region| {
@@ -199,12 +203,26 @@ impl<S: Spec<Fp, WIDTH, RATE> + Sync, const WIDTH: usize, const RATE: usize, con
                                 ValType::PrevAssigned(v) | ValType::AssignedConstant(v, ..) => {
                                     Ok(v.clone())
                                 }
-                                ValType::Constant(f) => region.assign_advice_from_constant(
-                                    || format!("load message_{}", i),
-                                    self.config.hash_inputs[x],
-                                    y,
-                                    *f,
-                                ),
+                                ValType::Constant(f) => {
+                                    if local_constants.contains_key(f) {
+                                        Ok(constants.get(f).unwrap().assigned_cell().ok_or({
+                                            log::error!("constant not previously assigned");
+                                            Error::Synthesis
+                                        })?)
+                                    } else {
+                                        let res = region.assign_advice_from_constant(
+                                            || format!("load message_{}", i),
+                                            self.config.hash_inputs[x],
+                                            y,
+                                            *f,
+                                        )?;
+
+                                        constants
+                                            .insert(*f, ValType::AssignedConstant(res.clone(), *f));
+
+                                        Ok(res)
+                                    }
+                                }
                                 e => {
                                     log::error!(
                                         "wrong input type {:?}, must be previously assigned",
@@ -270,8 +288,9 @@ impl<S: Spec<Fp, WIDTH, RATE> + Sync, const WIDTH: usize, const RATE: usize, con
         layouter: &mut impl Layouter<Fp>,
         input: &[ValTensor<Fp>],
         row_offset: usize,
+        constants: &mut ConstantsMap<Fp>,
     ) -> Result<ValTensor<Fp>, Error> {
-        let (mut input_cells, zero_val) = self.layout_inputs(layouter, input)?;
+        let (mut input_cells, zero_val) = self.layout_inputs(layouter, input, constants)?;
         // extract the values from the input cells
         let mut assigned_input: Tensor<ValType<Fp>> =
             input_cells.iter().map(|e| ValType::from(e.clone())).into();
@@ -434,7 +453,7 @@ mod tests {
         *,
     };
 
-    use std::marker::PhantomData;
+    use std::{collections::HashMap, marker::PhantomData};
 
     use halo2_gadgets::poseidon::primitives::Spec;
     use halo2_proofs::{
@@ -477,7 +496,12 @@ mod tests {
             mut layouter: impl Layouter<Fp>,
         ) -> Result<(), Error> {
             let chip: PoseidonChip<PoseidonSpec, WIDTH, RATE, L> = PoseidonChip::new(config);
-            chip.layout(&mut layouter, &[self.message.clone()], 0)?;
+            chip.layout(
+                &mut layouter,
+                &[self.message.clone()],
+                0,
+                &mut HashMap::new(),
+            )?;
 
             Ok(())
         }

src/circuit/ops/chip.rs

@@ -345,7 +345,7 @@ pub struct BaseConfig<F: PrimeField + TensorType + PartialOrd> {
     _marker: PhantomData<F>,
 }
 
-impl<F: PrimeField + TensorType + PartialOrd> BaseConfig<F> {
+impl<F: PrimeField + TensorType + PartialOrd + std::hash::Hash> BaseConfig<F> {
     /// Returns a new [BaseConfig] with no inputs, no selectors, and no tables.
     pub fn dummy(col_size: usize, num_inner_cols: usize) -> Self {
         Self {
@@ -956,20 +956,6 @@ impl<F: PrimeField + TensorType + PartialOrd> BaseConfig<F> {
         values: &[ValTensor<F>],
         op: Box<dyn Op<F>>,
     ) -> Result<Option<ValTensor<F>>, Box<dyn Error>> {
-        let res = op.layout(self, region, values)?;
-
-        if matches!(&self.check_mode, CheckMode::SAFE) && !region.is_dummy() {
-            if let Some(claimed_output) = &res {
-                // during key generation this will be unknown vals so we use this as a flag to check
-                let mut is_assigned = !claimed_output.any_unknowns()?;
-                for val in values.iter() {
-                    is_assigned = is_assigned && !val.any_unknowns()?;
-                }
-                if is_assigned {
-                    op.safe_mode_check(claimed_output, values)?;
-                }
-            }
-        };
-        Ok(res)
+        op.layout(self, region, values)
     }
 }

src/circuit/ops/hybrid.rs

@@ -1,9 +1,9 @@
 use super::*;
 use crate::{
     circuit::{layouts, utils, Tolerance},
-    fieldutils::{felt_to_i128, i128_to_felt},
+    fieldutils::i128_to_felt,
     graph::multiplier_to_scale,
-    tensor::{self, Tensor, TensorError, TensorType, ValTensor},
+    tensor::{self, Tensor, TensorType, ValTensor},
 };
 use halo2curves::ff::PrimeField;
 use serde::{Deserialize, Serialize};
@@ -29,15 +29,15 @@ pub enum HybridOp {
         dim: usize,
     },
     SumPool {
-        padding: [(usize, usize); 2],
-        stride: (usize, usize),
-        kernel_shape: (usize, usize),
+        padding: Vec<(usize, usize)>,
+        stride: Vec<usize>,
+        kernel_shape: Vec<usize>,
         normalized: bool,
     },
-    MaxPool2d {
-        padding: [(usize, usize); 2],
-        stride: (usize, usize),
-        pool_dims: (usize, usize),
+    MaxPool {
+        padding: Vec<(usize, usize)>,
+        stride: Vec<usize>,
+        pool_dims: Vec<usize>,
     },
     ReduceMin {
         axes: Vec<usize>,
@@ -46,7 +46,8 @@ pub enum HybridOp {
         dim: usize,
     },
     Softmax {
-        scale: utils::F32,
+        input_scale: utils::F32,
+        output_scale: utils::F32,
         axes: Vec<usize>,
     },
     RangeCheck(Tolerance),
@@ -70,7 +71,7 @@ pub enum HybridOp {
     },
 }
 
-impl<F: PrimeField + TensorType + PartialOrd> Op<F> for HybridOp {
+impl<F: PrimeField + TensorType + PartialOrd + std::hash::Hash> Op<F> for HybridOp {
     ///
     fn requires_homogenous_input_scales(&self) -> Vec<usize> {
         match self {
@@ -84,86 +85,6 @@ impl<F: PrimeField + TensorType + PartialOrd> Op<F> for HybridOp {
     fn as_any(&self) -> &dyn Any {
         self
     }
-    /// Matches a [Op] to an operation in the `tensor::ops` module.
-    fn f(&self, inputs: &[Tensor<F>]) -> Result<ForwardResult<F>, TensorError> {
-        let x = inputs[0].clone().map(|x| felt_to_i128(x));
-
-        let res = match &self {
-            HybridOp::ReduceMax { axes, .. } => tensor::ops::max_axes(&x, axes)?,
-            HybridOp::ReduceMin { axes, .. } => tensor::ops::min_axes(&x, axes)?,
-            HybridOp::Div { denom, .. } => {
-                crate::tensor::ops::nonlinearities::const_div(&x, denom.0 as f64)
-            }
-            HybridOp::Recip {
-                input_scale,
-                output_scale,
-                ..
-            } => crate::tensor::ops::nonlinearities::recip(
-                &x,
-                input_scale.0 as f64,
-                output_scale.0 as f64,
-            ),
-            HybridOp::ReduceArgMax { dim } => tensor::ops::argmax_axes(&x, *dim)?,
-            HybridOp::ReduceArgMin { dim } => tensor::ops::argmin_axes(&x, *dim)?,
-            HybridOp::Gather { dim, constant_idx } => {
-                if let Some(idx) = constant_idx {
-                    tensor::ops::gather(&x, idx, *dim)?
-                } else {
-                    let y = inputs[1].clone().map(|x| felt_to_i128(x));
-                    tensor::ops::gather(&x, &y.map(|x| x as usize), *dim)?
-                }
-            }
-            HybridOp::OneHot { dim, num_classes } => {
-                tensor::ops::one_hot(&x, *num_classes, *dim)?.clone()
-            }
-
-            HybridOp::TopK { dim, k, largest } => tensor::ops::topk_axes(&x, *k, *dim, *largest)?,
-            HybridOp::MaxPool2d {
-                padding,
-                stride,
-                pool_dims,
-                ..
-            } => tensor::ops::max_pool2d(&x, padding, stride, pool_dims)?,
-            HybridOp::SumPool {
-                padding,
-                stride,
-                kernel_shape,
-                normalized,
-            } => tensor::ops::sumpool(&x, *padding, *stride, *kernel_shape, *normalized)?,
-            HybridOp::Softmax { scale, axes } => {
-                tensor::ops::nonlinearities::softmax_axes(&x, scale.into(), axes)
-            }
-            HybridOp::RangeCheck(tol) => {
-                let y = inputs[1].clone().map(|x| felt_to_i128(x));
-                tensor::ops::nonlinearities::range_check_percent(&[x, y], 128, 128, tol.val)
-            }
-            HybridOp::Greater => {
-                let y = inputs[1].clone().map(|x| felt_to_i128(x));
-                tensor::ops::greater(&x, &y)?
-            }
-            HybridOp::GreaterEqual => {
-                let y = inputs[1].clone().map(|x| felt_to_i128(x));
-                tensor::ops::greater_equal(&x, &y)?
-            }
-            HybridOp::Less => {
-                let y = inputs[1].clone().map(|x| felt_to_i128(x));
-                tensor::ops::less(&x, &y)?
-            }
-            HybridOp::LessEqual => {
-                let y = inputs[1].clone().map(|x| felt_to_i128(x));
-                tensor::ops::less_equal(&x, &y)?
-            }
-            HybridOp::Equals => {
-                let y = inputs[1].clone().map(|x| felt_to_i128(x));
-                tensor::ops::equals(&x, &y)?
-            }
-        };
-
-        // convert back to felt
-        let output = res.map(|x| i128_to_felt(x));
-
-        Ok(ForwardResult { output })
-    }
 
     fn as_string(&self) -> String {
         match self {
@@ -193,18 +114,25 @@ impl<F: PrimeField + TensorType + PartialOrd> Op<F> for HybridOp {
             ),
             HybridOp::ReduceMax { axes } => format!("REDUCEMAX (axes={:?})", axes),
             HybridOp::ReduceArgMax { dim } => format!("REDUCEARGMAX (dim={})", dim),
-            HybridOp::MaxPool2d {
+            HybridOp::MaxPool {
                 padding,
                 stride,
                 pool_dims,
             } => format!(
-                "MAXPOOL2D (padding={:?}, stride={:?}, pool_dims={:?})",
+                "MaxPool (padding={:?}, stride={:?}, pool_dims={:?})",
                 padding, stride, pool_dims
             ),
             HybridOp::ReduceMin { axes } => format!("REDUCEMIN (axes={:?})", axes),
             HybridOp::ReduceArgMin { dim } => format!("REDUCEARGMIN (dim={})", dim),
-            HybridOp::Softmax { scale, axes } => {
-                format!("SOFTMAX (scale={}, axes={:?})", scale, axes)
+            HybridOp::Softmax {
+                input_scale,
+                output_scale,
+                axes,
+            } => {
+                format!(
+                    "SOFTMAX (input_scale={}, output_scale={}, axes={:?})",
+                    input_scale, output_scale, axes
+                )
             }
             HybridOp::RangeCheck(p) => format!("RANGECHECK (tol={:?})", p),
             HybridOp::Greater => "GREATER".into(),
@@ -238,9 +166,9 @@ impl<F: PrimeField + TensorType + PartialOrd> Op<F> for HybridOp {
                 config,
                 region,
                 values[..].try_into()?,
-                *padding,
-                *stride,
-                *kernel_shape,
+                padding,
+                stride,
+                kernel_shape,
                 *normalized,
             )?,
             HybridOp::Recip {
@@ -300,17 +228,17 @@ impl<F: PrimeField + TensorType + PartialOrd> Op<F> for HybridOp {
                 }
             }
 
-            HybridOp::MaxPool2d {
+            HybridOp::MaxPool {
                 padding,
                 stride,
                 pool_dims,
-            } => layouts::max_pool2d(
+            } => layouts::max_pool(
                 config,
                 region,
                 values[..].try_into()?,
-                *padding,
-                *stride,
-                *pool_dims,
+                padding,
+                stride,
+                pool_dims,
             )?,
             HybridOp::ReduceMax { axes } => {
                 layouts::max_axes(config, region, values[..].try_into()?, axes)?
@@ -324,9 +252,18 @@ impl<F: PrimeField + TensorType + PartialOrd> Op<F> for HybridOp {
             HybridOp::ReduceArgMin { dim } => {
                 layouts::argmin_axes(config, region, values[..].try_into()?, *dim)?
             }
-            HybridOp::Softmax { scale, axes } => {
-                layouts::softmax_axes(config, region, values[..].try_into()?, *scale, axes)?
-            }
+            HybridOp::Softmax {
+                input_scale,
+                output_scale,
+                axes,
+            } => layouts::softmax_axes(
+                config,
+                region,
+                values[..].try_into()?,
+                *input_scale,
+                *output_scale,
+                axes,
+            )?,
             HybridOp::RangeCheck(tol) => layouts::range_check_percent(
                 config,
                 region,
@@ -359,8 +296,9 @@ impl<F: PrimeField + TensorType + PartialOrd> Op<F> for HybridOp {
             | HybridOp::ReduceArgMax { .. }
             | HybridOp::OneHot { .. }
             | HybridOp::ReduceArgMin { .. } => 0,
-            HybridOp::Softmax { .. } => 2 * in_scales[0],
-            HybridOp::Recip { output_scale, .. } => multiplier_to_scale(output_scale.0 as f64),
+            HybridOp::Softmax { output_scale, .. } | HybridOp::Recip { output_scale, .. } => {
+                multiplier_to_scale(output_scale.0 as f64)
+            }
             _ => in_scales[0],
         };
         Ok(scale)

src/circuit/ops/lookup.rs

@@ -123,6 +123,9 @@ pub enum LookupOp {
         scale: utils::F32,
         a: utils::F32,
     },
+    HardSwish {
+        scale: utils::F32,
+    },
 }
 
 impl LookupOp {
@@ -132,15 +135,12 @@ impl LookupOp {
         let range = range as i128;
         (-range, range)
     }
-}
 
-impl<F: PrimeField + TensorType + PartialOrd> Op<F> for LookupOp {
-    /// Returns a reference to the Any trait.
-    fn as_any(&self) -> &dyn Any {
-        self
-    }
     /// Matches a [Op] to an operation in the `tensor::ops` module.
-    fn f(&self, x: &[Tensor<F>]) -> Result<ForwardResult<F>, TensorError> {
+    pub(crate) fn f<F: PrimeField + TensorType + PartialOrd + std::hash::Hash>(
+        &self,
+        x: &[Tensor<F>],
+    ) -> Result<ForwardResult<F>, TensorError> {
         let x = x[0].clone().map(|x| felt_to_i128(x));
         let res = match &self {
             LookupOp::Abs => Ok(tensor::ops::abs(&x)?),
@@ -223,12 +223,22 @@ impl<F: PrimeField + TensorType + PartialOrd> Op<F> for LookupOp {
             LookupOp::ATan { scale } => Ok(tensor::ops::nonlinearities::atan(&x, scale.into())),
             LookupOp::ATanh { scale } => Ok(tensor::ops::nonlinearities::atanh(&x, scale.into())),
             LookupOp::Tanh { scale } => Ok(tensor::ops::nonlinearities::tanh(&x, scale.into())),
+            LookupOp::HardSwish { scale } => {
+                Ok(tensor::ops::nonlinearities::hardswish(&x, scale.into()))
+            }
         }?;
 
         let output = res.map(|x| i128_to_felt(x));
 
         Ok(ForwardResult { output })
     }
+}
+
+impl<F: PrimeField + TensorType + PartialOrd + std::hash::Hash> Op<F> for LookupOp {
+    /// Returns a reference to the Any trait.
+    fn as_any(&self) -> &dyn Any {
+        self
+    }
 
     /// Returns the name of the operation
     fn as_string(&self) -> String {
@@ -276,6 +286,7 @@ impl<F: PrimeField + TensorType + PartialOrd> Op<F> for LookupOp {
             LookupOp::ASin { scale } => format!("ASIN(scale={})", scale),
             LookupOp::Sinh { scale } => format!("SINH(scale={})", scale),
             LookupOp::ASinh { scale } => format!("ASINH(scale={})", scale),
+            LookupOp::HardSwish { scale } => format!("HARDSWISH(scale={})", scale),
         }
     }
 

src/circuit/ops/mod.rs

@@ -4,7 +4,7 @@ use serde::{Deserialize, Serialize};
 
 use crate::{
     graph::quantize_tensor,
-    tensor::{self, Tensor, TensorError, TensorType, ValTensor},
+    tensor::{self, Tensor, TensorType, ValTensor},
 };
 use halo2curves::ff::PrimeField;
 
@@ -27,14 +27,14 @@ pub mod region;
 
 /// A struct representing the result of a forward pass.
 #[derive(Clone, Debug, PartialEq, Eq, PartialOrd, Ord, Serialize, Deserialize)]
-pub struct ForwardResult<F: PrimeField + TensorType + PartialOrd> {
+pub struct ForwardResult<F: PrimeField + TensorType + PartialOrd + std::hash::Hash> {
     pub(crate) output: Tensor<F>,
 }
 
 /// A trait representing operations that can be represented as constraints in a circuit.
-pub trait Op<F: PrimeField + TensorType + PartialOrd>: std::fmt::Debug + Send + Sync + Any {
-    /// Matches a [Op] to an operation in the `tensor::ops` module.
-    fn f(&self, x: &[Tensor<F>]) -> Result<ForwardResult<F>, TensorError>;
+pub trait Op<F: PrimeField + TensorType + PartialOrd + std::hash::Hash>:
+    std::fmt::Debug + Send + Sync + Any
+{
     /// Returns a string representation of the operation.
     fn as_string(&self) -> String;
 
@@ -69,36 +69,9 @@ pub trait Op<F: PrimeField + TensorType + PartialOrd>: std::fmt::Debug + Send +
 
     /// Returns a reference to the Any trait.
     fn as_any(&self) -> &dyn Any;
-
-    /// Safe mode output checl
-    fn safe_mode_check(
-        &self,
-        claimed_output: &ValTensor<F>,
-        original_values: &[ValTensor<F>],
-    ) -> Result<(), TensorError> {
-        let felt_evals = original_values
-            .iter()
-            .map(|v| {
-                let mut evals = v.get_felt_evals().map_err(|_| TensorError::FeltError)?;
-                evals.reshape(v.dims())?;
-                Ok(evals)
-            })
-            .collect::<Result<Vec<_>, _>>()?;
-
-        let ref_op: Tensor<F> = self.f(&felt_evals)?.output;
-
-        let mut output = claimed_output
-            .get_felt_evals()
-            .map_err(|_| TensorError::FeltError)?;
-        output.reshape(claimed_output.dims())?;
-
-        assert_eq!(output, ref_op);
-
-        Ok(())
-    }
 }
 
-impl<F: PrimeField + TensorType + PartialOrd> Clone for Box<dyn Op<F>> {
+impl<F: PrimeField + TensorType + PartialOrd + std::hash::Hash> Clone for Box<dyn Op<F>> {
     fn clone(&self) -> Self {
         self.clone_dyn()
     }
@@ -165,7 +138,7 @@ pub struct Input {
     pub datum_type: InputType,
 }
 
-impl<F: PrimeField + TensorType + PartialOrd> Op<F> for Input {
+impl<F: PrimeField + TensorType + PartialOrd + std::hash::Hash> Op<F> for Input {
     fn out_scale(&self, _: Vec<crate::Scale>) -> Result<crate::Scale, Box<dyn Error>> {
         Ok(self.scale)
     }
@@ -174,12 +147,6 @@ impl<F: PrimeField + TensorType + PartialOrd> Op<F> for Input {
         self
     }
 
-    fn f(&self, x: &[Tensor<F>]) -> Result<ForwardResult<F>, TensorError> {
-        Ok(ForwardResult {
-            output: x[0].clone(),
-        })
-    }
-
     fn as_string(&self) -> String {
         "Input".into()
     }
@@ -226,16 +193,13 @@ impl<F: PrimeField + TensorType + PartialOrd> Op<F> for Input {
 #[derive(Clone, Debug, PartialEq, Eq, Hash, PartialOrd, Ord, Serialize, Deserialize)]
 pub struct Unknown;
 
-impl<F: PrimeField + TensorType + PartialOrd> Op<F> for Unknown {
+impl<F: PrimeField + TensorType + PartialOrd + std::hash::Hash> Op<F> for Unknown {
     fn out_scale(&self, _: Vec<crate::Scale>) -> Result<crate::Scale, Box<dyn Error>> {
         Ok(0)
     }
     fn as_any(&self) -> &dyn Any {
         self
     }
-    fn f(&self, _: &[Tensor<F>]) -> Result<ForwardResult<F>, TensorError> {
-        Err(TensorError::WrongMethod)
-    }
 
     fn as_string(&self) -> String {
         "Unknown".into()
@@ -256,7 +220,7 @@ impl<F: PrimeField + TensorType + PartialOrd> Op<F> for Unknown {
 
 ///
 #[derive(Clone, Debug, Serialize, Deserialize)]
-pub struct Constant<F: PrimeField + TensorType + PartialOrd> {
+pub struct Constant<F: PrimeField + TensorType + PartialOrd + std::hash::Hash> {
     ///
     pub quantized_values: Tensor<F>,
     ///
@@ -266,7 +230,7 @@ pub struct Constant<F: PrimeField + TensorType + PartialOrd> {
     pub pre_assigned_val: Option<ValTensor<F>>,
 }
 
-impl<F: PrimeField + TensorType + PartialOrd> Constant<F> {
+impl<F: PrimeField + TensorType + PartialOrd + std::hash::Hash> Constant<F> {
     ///
     pub fn new(quantized_values: Tensor<F>, raw_values: Tensor<f32>) -> Self {
         Self {
@@ -293,17 +257,18 @@ impl<F: PrimeField + TensorType + PartialOrd> Constant<F> {
     }
 }
 
-impl<F: PrimeField + TensorType + PartialOrd + Serialize + for<'de> Deserialize<'de>> Op<F>
-    for Constant<F>
+impl<
+        F: PrimeField
+            + TensorType
+            + PartialOrd
+            + std::hash::Hash
+            + Serialize
+            + for<'de> Deserialize<'de>,
+    > Op<F> for Constant<F>
 {
     fn as_any(&self) -> &dyn Any {
         self
     }
-    fn f(&self, _: &[Tensor<F>]) -> Result<ForwardResult<F>, TensorError> {
-        let output = self.quantized_values.clone();
-
-        Ok(ForwardResult { output })
-    }
 
     fn as_string(&self) -> String {
         format!("CONST (scale={})", self.quantized_values.scale().unwrap())

src/circuit/ops/poly.rs

@@ -1,6 +1,5 @@
 use crate::{
     circuit::layouts,
-    fieldutils::felt_to_i128,
     tensor::{self, Tensor, TensorError},
 };
 
@@ -32,8 +31,8 @@ pub enum PolyOp {
         equation: String,
     },
     Conv {
-        padding: [(usize, usize); 2],
-        stride: (usize, usize),
+        padding: Vec<(usize, usize)>,
+        stride: Vec<usize>,
     },
     Downsample {
         axis: usize,
@@ -41,9 +40,9 @@ pub enum PolyOp {
         modulo: usize,
     },
     DeConv {
-        padding: [(usize, usize); 2],
-        output_padding: (usize, usize),
-        stride: (usize, usize),
+        padding: Vec<(usize, usize)>,
+        output_padding: Vec<usize>,
+        stride: Vec<usize>,
     },
     Add,
     Sub,
@@ -58,10 +57,13 @@ pub enum PolyOp {
         destination: usize,
     },
     Flatten(Vec<usize>),
-    Pad([(usize, usize); 2]),
+    Pad(Vec<(usize, usize)>),
     Sum {
         axes: Vec<usize>,
     },
+    MeanOfSquares {
+        axes: Vec<usize>,
+    },
     Prod {
         axes: Vec<usize>,
         len_prod: usize,
@@ -83,10 +85,20 @@ pub enum PolyOp {
     And,
     Or,
     Xor,
+    Trilu {
+        upper: bool,
+        k: i32,
+    },
 }
 
-impl<F: PrimeField + TensorType + PartialOrd + Serialize + for<'de> Deserialize<'de>> Op<F>
-    for PolyOp
+impl<
+        F: PrimeField
+            + TensorType
+            + PartialOrd
+            + std::hash::Hash
+            + Serialize
+            + for<'de> Deserialize<'de>,
+    > Op<F> for PolyOp
 {
     /// Returns a reference to the Any trait.
     fn as_any(&self) -> &dyn Any {
@@ -95,10 +107,28 @@ impl<F: PrimeField + TensorType + PartialOrd + Serialize + for<'de> Deserialize<
 
     fn as_string(&self) -> String {
         match &self {
-            PolyOp::GatherElements { dim, .. } => format!("GATHERELEMENTS (dim={})", dim),
-            PolyOp::GatherND { batch_dims, .. } => format!("GATHERND (batch_dims={})", batch_dims),
-            PolyOp::ScatterElements { dim, .. } => format!("SCATTERELEMENTS (dim={})", dim),
-            PolyOp::ScatterND { .. } => "SCATTERND".into(),
+            PolyOp::GatherElements { dim, constant_idx } => format!(
+                "GATHERELEMENTS (dim={}, constant_idx{})",
+                dim,
+                constant_idx.is_some()
+            ),
+            PolyOp::GatherND {
+                batch_dims,
+                indices,
+            } => format!(
+                "GATHERND (batch_dims={}, constant_idx{})",
+                batch_dims,
+                indices.is_some()
+            ),
+            PolyOp::MeanOfSquares { axes } => format!("MEANOFSQUARES (axes={:?})", axes),
+            PolyOp::ScatterElements { dim, constant_idx } => format!(
+                "SCATTERELEMENTS (dim={}, constant_idx{})",
+                dim,
+                constant_idx.is_some()
+            ),
+            PolyOp::ScatterND { constant_idx } => {
+                format!("SCATTERND (constant_idx={})", constant_idx.is_some())
+            }
             PolyOp::MultiBroadcastTo { shape } => format!("MULTIBROADCASTTO (shape={:?})", shape),
             PolyOp::MoveAxis { .. } => "MOVEAXIS".into(),
             PolyOp::Downsample { .. } => "DOWNSAMPLE".into(),
@@ -110,15 +140,26 @@ impl<F: PrimeField + TensorType + PartialOrd + Serialize + for<'de> Deserialize<
             }
             PolyOp::Reshape(shape) => format!("RESHAPE (shape={:?})", shape),
             PolyOp::Flatten(_) => "FLATTEN".into(),
-            PolyOp::Pad(_) => "PAD".into(),
+            PolyOp::Pad(pads) => format!("PAD (pads={:?})", pads),
             PolyOp::Add => "ADD".into(),
             PolyOp::Mult => "MULT".into(),
             PolyOp::Sub => "SUB".into(),
-            PolyOp::Sum { .. } => "SUM".into(),
+            PolyOp::Sum { axes } => format!("SUM (axes={:?})", axes),
             PolyOp::Prod { .. } => "PROD".into(),
             PolyOp::Pow(_) => "POW".into(),
-            PolyOp::Conv { .. } => "CONV".into(),
-            PolyOp::DeConv { .. } => "DECONV".into(),
+            PolyOp::Conv { stride, padding } => {
+                format!("CONV (stride={:?}, padding={:?})", stride, padding)
+            }
+            PolyOp::DeConv {
+                stride,
+                padding,
+                output_padding,
+            } => {
+                format!(
+                    "DECONV (stride={:?}, padding={:?}, output_padding={:?})",
+                    stride, padding, output_padding
+                )
+            }
             PolyOp::Concat { axis } => format!("CONCAT (axis={})", axis),
             PolyOp::Slice { axis, start, end } => {
                 format!("SLICE (axis={}, start={}, end={})", axis, start, end)
@@ -128,148 +169,10 @@ impl<F: PrimeField + TensorType + PartialOrd + Serialize + for<'de> Deserialize<
             PolyOp::And => "AND".into(),
             PolyOp::Or => "OR".into(),
             PolyOp::Xor => "XOR".into(),
+            PolyOp::Trilu { upper, k } => format!("TRILU (upper={}, k={})", upper, k),
         }
     }
 
-    /// Matches a [Op] to an operation in the `tensor::ops` module.
-    fn f(&self, inputs: &[Tensor<F>]) -> Result<ForwardResult<F>, TensorError> {
-        let mut inputs = inputs.to_vec();
-        let res = match &self {
-            PolyOp::MultiBroadcastTo { shape } => {
-                if 1 != inputs.len() {
-                    return Err(TensorError::DimMismatch(
-                        "multibroadcastto inputs".to_string(),
-                    ));
-                }
-                inputs[0].expand(shape)
-            }
-            PolyOp::And => tensor::ops::and(&inputs[0], &inputs[1]),
-            PolyOp::Or => tensor::ops::or(&inputs[0], &inputs[1]),
-            PolyOp::Xor => tensor::ops::xor(&inputs[0], &inputs[1]),
-            PolyOp::Not => tensor::ops::not(&inputs[0]),
-            PolyOp::Downsample {
-                axis,
-                stride,
-                modulo,
-            } => tensor::ops::downsample(&inputs[0], *axis, *stride, *modulo),
-            PolyOp::Resize { scale_factor } => tensor::ops::resize(&inputs[0], scale_factor),
-            PolyOp::Iff => tensor::ops::iff(&inputs[0], &inputs[1], &inputs[2]),
-            PolyOp::Einsum { equation } => tensor::ops::einsum(equation, &inputs),
-            PolyOp::Identity { .. } => Ok(inputs[0].clone()),
-            PolyOp::Reshape(new_dims) => {
-                let mut t = inputs[0].clone();
-                t.reshape(new_dims)?;
-                Ok(t)
-            }
-            PolyOp::MoveAxis {
-                source,
-                destination,
-            } => inputs[0].move_axis(*source, *destination),
-            PolyOp::Flatten(new_dims) => {
-                let mut t = inputs[0].clone();
-                t.reshape(new_dims)?;
-                Ok(t)
-            }
-            PolyOp::Pad(p) => {
-                if 1 != inputs.len() {
-                    return Err(TensorError::DimMismatch("pad inputs".to_string()));
-                }
-                tensor::ops::pad(&inputs[0], *p)
-            }
-            PolyOp::Add => tensor::ops::add(&inputs),
-            PolyOp::Neg => tensor::ops::neg(&inputs[0]),
-            PolyOp::Sub => tensor::ops::sub(&inputs),
-            PolyOp::Mult => tensor::ops::mult(&inputs),
-            PolyOp::Conv { padding, stride } => tensor::ops::conv(&inputs, *padding, *stride),
-            PolyOp::DeConv {
-                padding,
-                output_padding,
-                stride,
-            } => tensor::ops::deconv(&inputs, *padding, *output_padding, *stride),
-            PolyOp::Pow(u) => {
-                if 1 != inputs.len() {
-                    return Err(TensorError::DimMismatch("pow inputs".to_string()));
-                }
-                inputs[0].pow(*u)
-            }
-            PolyOp::Sum { axes } => {
-                if 1 != inputs.len() {
-                    return Err(TensorError::DimMismatch("sum inputs".to_string()));
-                }
-                tensor::ops::sum_axes(&inputs[0], axes)
-            }
-            PolyOp::Prod { axes, .. } => {
-                if 1 != inputs.len() {
-                    return Err(TensorError::DimMismatch("prod inputs".to_string()));
-                }
-                tensor::ops::prod_axes(&inputs[0], axes)
-            }
-            PolyOp::Concat { axis } => {
-                tensor::ops::concat(&inputs.iter().collect::<Vec<_>>(), *axis)
-            }
-            PolyOp::Slice { axis, start, end } => {
-                if 1 != inputs.len() {
-                    return Err(TensorError::DimMismatch("slice inputs".to_string()));
-                }
-                tensor::ops::slice(&inputs[0], axis, start, end)
-            }
-            PolyOp::GatherElements { dim, constant_idx } => {
-                let x = inputs[0].clone();
-                let y = if let Some(idx) = constant_idx {
-                    idx.clone()
-                } else {
-                    inputs[1].clone().map(|x| felt_to_i128(x) as usize)
-                };
-                tensor::ops::gather_elements(&x, &y, *dim)
-            }
-            PolyOp::GatherND {
-                indices,
-                batch_dims,
-            } => {
-                let x = inputs[0].clone();
-                let y = if let Some(idx) = indices {
-                    idx.clone()
-                } else {
-                    inputs[1].clone().map(|x| felt_to_i128(x) as usize)
-                };
-                tensor::ops::gather_nd(&x, &y, *batch_dims)
-            }
-            PolyOp::ScatterElements { dim, constant_idx } => {
-                let x = inputs[0].clone();
-
-                let idx = if let Some(idx) = constant_idx {
-                    idx.clone()
-                } else {
-                    inputs[1].clone().map(|x| felt_to_i128(x) as usize)
-                };
-
-                let src = if constant_idx.is_some() {
-                    inputs[1].clone()
-                } else {
-                    inputs[2].clone()
-                };
-                tensor::ops::scatter(&x, &idx, &src, *dim)
-            }
-
-            PolyOp::ScatterND { constant_idx } => {
-                let x = inputs[0].clone();
-                let idx = if let Some(idx) = constant_idx {
-                    idx.clone()
-                } else {
-                    inputs[1].clone().map(|x| felt_to_i128(x) as usize)
-                };
-                let src = if constant_idx.is_some() {
-                    inputs[1].clone()
-                } else {
-                    inputs[2].clone()
-                };
-                tensor::ops::scatter_nd(&x, &idx, &src)
-            }
-        }?;
-
-        Ok(ForwardResult { output: res })
-    }
-
     fn layout(
         &self,
         config: &mut crate::circuit::BaseConfig<F>,
@@ -280,6 +183,9 @@ impl<F: PrimeField + TensorType + PartialOrd + Serialize + for<'de> Deserialize<
             PolyOp::MultiBroadcastTo { shape } => {
                 layouts::expand(config, region, values[..].try_into()?, shape)?
             }
+            PolyOp::MeanOfSquares { axes } => {
+                layouts::mean_of_squares_axes(config, region, values[..].try_into()?, axes)?
+            }
             PolyOp::Xor => layouts::xor(config, region, values[..].try_into()?)?,
             PolyOp::Or => layouts::or(config, region, values[..].try_into()?)?,
             PolyOp::And => layouts::and(config, region, values[..].try_into()?)?,
@@ -306,7 +212,7 @@ impl<F: PrimeField + TensorType + PartialOrd + Serialize + for<'de> Deserialize<
                 layouts::prod_axes(config, region, values[..].try_into()?, axes)?
             }
             PolyOp::Conv { padding, stride } => {
-                layouts::conv(config, region, values[..].try_into()?, *padding, *stride)?
+                layouts::conv(config, region, values[..].try_into()?, padding, stride)?
             }
             PolyOp::GatherElements { dim, constant_idx } => {
                 if let Some(idx) = constant_idx {
@@ -358,9 +264,9 @@ impl<F: PrimeField + TensorType + PartialOrd + Serialize + for<'de> Deserialize<
                 config,
                 region,
                 values[..].try_into()?,
-                *padding,
-                *output_padding,
-                *stride,
+                padding,
+                output_padding,
+                stride,
             )?,
             PolyOp::Add => layouts::pairwise(config, region, values[..].try_into()?, BaseOp::Add)?,
             PolyOp::Sub => layouts::pairwise(config, region, values[..].try_into()?, BaseOp::Sub)?,
@@ -376,7 +282,7 @@ impl<F: PrimeField + TensorType + PartialOrd + Serialize + for<'de> Deserialize<
                     )));
                 }
                 let mut input = values[0].clone();
-                input.pad(*p)?;
+                input.pad(p.clone(), 0)?;
                 input
             }
             PolyOp::Pow(exp) => layouts::pow(config, region, values[..].try_into()?, *exp)?,
@@ -384,11 +290,15 @@ impl<F: PrimeField + TensorType + PartialOrd + Serialize + for<'de> Deserialize<
             PolyOp::Slice { axis, start, end } => {
                 layouts::slice(config, region, values[..].try_into()?, axis, start, end)?
             }
+            PolyOp::Trilu { upper, k } => {
+                layouts::trilu(config, region, values[..].try_into()?, k, upper)?
+            }
         }))
     }
 
     fn out_scale(&self, in_scales: Vec<crate::Scale>) -> Result<crate::Scale, Box<dyn Error>> {
         let scale = match self {
+            PolyOp::MeanOfSquares { .. } => 2 * in_scales[0],
             PolyOp::Xor | PolyOp::Or | PolyOp::And | PolyOp::Not => 0,
             PolyOp::Iff => in_scales[1],
             PolyOp::Einsum { .. } => {

src/circuit/ops/region.rs

@@ -2,24 +2,28 @@ use crate::{
     circuit::table::Range,
     tensor::{Tensor, TensorError, TensorType, ValTensor, ValType, VarTensor},
 };
+#[cfg(not(target_arch = "wasm32"))]
+use colored::Colorize;
 use halo2_proofs::{
     circuit::Region,
     plonk::{Error, Selector},
 };
 use halo2curves::ff::PrimeField;
+use portable_atomic::AtomicI128 as AtomicInt;
 use std::{
     cell::RefCell,
-    collections::HashSet,
+    collections::{HashMap, HashSet},
     sync::{
         atomic::{AtomicUsize, Ordering},
         Arc, Mutex,
     },
 };
 
-use portable_atomic::AtomicI128 as AtomicInt;
-
 use super::lookup::LookupOp;
 
+/// Constants map
+pub type ConstantsMap<F> = HashMap<F, ValType<F>>;
+
 /// Dynamic lookup index
 #[derive(Clone, Debug, Default)]
 pub struct DynamicLookupIndex {
@@ -120,12 +124,11 @@ impl From<Box<dyn std::error::Error>> for RegionError {
 
 #[derive(Debug)]
 /// A context for a region
-pub struct RegionCtx<'a, F: PrimeField + TensorType + PartialOrd> {
+pub struct RegionCtx<'a, F: PrimeField + TensorType + PartialOrd + std::hash::Hash> {
     region: Option<RefCell<Region<'a, F>>>,
     row: usize,
     linear_coord: usize,
     num_inner_cols: usize,
-    total_constants: usize,
     dynamic_lookup_index: DynamicLookupIndex,
     shuffle_index: ShuffleIndex,
     used_lookups: HashSet<LookupOp>,
@@ -133,13 +136,34 @@ pub struct RegionCtx<'a, F: PrimeField + TensorType + PartialOrd> {
     max_lookup_inputs: i128,
     min_lookup_inputs: i128,
     max_range_size: i128,
-    throw_range_check_error: bool,
+    witness_gen: bool,
+    assigned_constants: ConstantsMap<F>,
 }
 
-impl<'a, F: PrimeField + TensorType + PartialOrd> RegionCtx<'a, F> {
+impl<'a, F: PrimeField + TensorType + PartialOrd + std::hash::Hash> RegionCtx<'a, F> {
+    #[cfg(not(target_arch = "wasm32"))]
     ///
-    pub fn increment_total_constants(&mut self, n: usize) {
-        self.total_constants += n;
+    pub fn debug_report(&self) {
+        log::debug!(
+            "(rows={}, coord={}, constants={}, max_lookup_inputs={}, min_lookup_inputs={}, max_range_size={}, dynamic_lookup_col_coord={}, shuffle_col_coord={})",
+            self.row().to_string().blue(),
+            self.linear_coord().to_string().yellow(),
+            self.total_constants().to_string().red(),
+            self.max_lookup_inputs().to_string().green(),
+            self.min_lookup_inputs().to_string().green(),
+            self.max_range_size().to_string().green(),
+            self.dynamic_lookup_col_coord().to_string().green(),
+            self.shuffle_col_coord().to_string().green());
+    }
+
+    ///
+    pub fn assigned_constants(&self) -> &ConstantsMap<F> {
+        &self.assigned_constants
+    }
+
+    ///
+    pub fn update_constants(&mut self, constants: ConstantsMap<F>) {
+        self.assigned_constants.extend(constants);
     }
 
     ///
@@ -163,8 +187,8 @@ impl<'a, F: PrimeField + TensorType + PartialOrd> RegionCtx<'a, F> {
     }
 
     ///
-    pub fn throw_range_check_error(&self) -> bool {
-        self.throw_range_check_error
+    pub fn witness_gen(&self) -> bool {
+        self.witness_gen
     }
 
     /// Create a new region context
@@ -177,7 +201,6 @@ impl<'a, F: PrimeField + TensorType + PartialOrd> RegionCtx<'a, F> {
             num_inner_cols,
             row,
             linear_coord,
-            total_constants: 0,
             dynamic_lookup_index: DynamicLookupIndex::default(),
             shuffle_index: ShuffleIndex::default(),
             used_lookups: HashSet::new(),
@@ -185,9 +208,22 @@ impl<'a, F: PrimeField + TensorType + PartialOrd> RegionCtx<'a, F> {
             max_lookup_inputs: 0,
             min_lookup_inputs: 0,
             max_range_size: 0,
-            throw_range_check_error: false,
+            witness_gen: true,
+            assigned_constants: HashMap::new(),
         }
     }
+
+    /// Create a new region context
+    pub fn new_with_constants(
+        region: Region<'a, F>,
+        row: usize,
+        num_inner_cols: usize,
+        constants: ConstantsMap<F>,
+    ) -> RegionCtx<'a, F> {
+        let mut new_self = Self::new(region, row, num_inner_cols);
+        new_self.assigned_constants = constants;
+        new_self
+    }
     /// Create a new region context from a wrapped region
     pub fn from_wrapped_region(
         region: Option<RefCell<Region<'a, F>>>,
@@ -202,7 +238,6 @@ impl<'a, F: PrimeField + TensorType + PartialOrd> RegionCtx<'a, F> {
             num_inner_cols,
             linear_coord,
             row,
-            total_constants: 0,
             dynamic_lookup_index,
             shuffle_index,
             used_lookups: HashSet::new(),
@@ -210,16 +245,13 @@ impl<'a, F: PrimeField + TensorType + PartialOrd> RegionCtx<'a, F> {
             max_lookup_inputs: 0,
             min_lookup_inputs: 0,
             max_range_size: 0,
-            throw_range_check_error: false,
+            witness_gen: false,
+            assigned_constants: HashMap::new(),
         }
     }
 
     /// Create a new region context
-    pub fn new_dummy(
-        row: usize,
-        num_inner_cols: usize,
-        throw_range_check_error: bool,
-    ) -> RegionCtx<'a, F> {
+    pub fn new_dummy(row: usize, num_inner_cols: usize, witness_gen: bool) -> RegionCtx<'a, F> {
         let region = None;
         let linear_coord = row * num_inner_cols;
 
@@ -228,7 +260,6 @@ impl<'a, F: PrimeField + TensorType + PartialOrd> RegionCtx<'a, F> {
             num_inner_cols,
             linear_coord,
             row,
-            total_constants: 0,
             dynamic_lookup_index: DynamicLookupIndex::default(),
             shuffle_index: ShuffleIndex::default(),
             used_lookups: HashSet::new(),
@@ -236,17 +267,17 @@ impl<'a, F: PrimeField + TensorType + PartialOrd> RegionCtx<'a, F> {
             max_lookup_inputs: 0,
             min_lookup_inputs: 0,
             max_range_size: 0,
-            throw_range_check_error,
+            witness_gen,
+            assigned_constants: HashMap::new(),
         }
     }
 
     /// Create a new region context
-    pub fn new_dummy_with_constants(
+    pub fn new_dummy_with_linear_coord(
         row: usize,
         linear_coord: usize,
-        total_constants: usize,
         num_inner_cols: usize,
-        throw_range_check_error: bool,
+        witness_gen: bool,
     ) -> RegionCtx<'a, F> {
         let region = None;
         RegionCtx {
@@ -254,7 +285,6 @@ impl<'a, F: PrimeField + TensorType + PartialOrd> RegionCtx<'a, F> {
             num_inner_cols,
             linear_coord,
             row,
-            total_constants,
             dynamic_lookup_index: DynamicLookupIndex::default(),
             shuffle_index: ShuffleIndex::default(),
             used_lookups: HashSet::new(),
@@ -262,7 +292,8 @@ impl<'a, F: PrimeField + TensorType + PartialOrd> RegionCtx<'a, F> {
             max_lookup_inputs: 0,
             min_lookup_inputs: 0,
             max_range_size: 0,
-            throw_range_check_error,
+            witness_gen,
+            assigned_constants: HashMap::new(),
         }
     }
 
@@ -312,29 +343,27 @@ impl<'a, F: PrimeField + TensorType + PartialOrd> RegionCtx<'a, F> {
     ) -> Result<(), RegionError> {
         let row = AtomicUsize::new(self.row());
         let linear_coord = AtomicUsize::new(self.linear_coord());
-        let constants = AtomicUsize::new(self.total_constants());
         let max_lookup_inputs = AtomicInt::new(self.max_lookup_inputs());
         let min_lookup_inputs = AtomicInt::new(self.min_lookup_inputs());
         let lookups = Arc::new(Mutex::new(self.used_lookups.clone()));
         let range_checks = Arc::new(Mutex::new(self.used_range_checks.clone()));
         let dynamic_lookup_index = Arc::new(Mutex::new(self.dynamic_lookup_index.clone()));
         let shuffle_index = Arc::new(Mutex::new(self.shuffle_index.clone()));
+        let constants = Arc::new(Mutex::new(self.assigned_constants.clone()));
 
         *output = output
             .par_enum_map(|idx, _| {
                 // we kick off the loop with the current offset
                 let starting_offset = row.load(Ordering::SeqCst);
                 let starting_linear_coord = linear_coord.load(Ordering::SeqCst);
-                let starting_constants = constants.load(Ordering::SeqCst);
                 // get inner value of the locked lookups
 
                 // we need to make sure that the region is not shared between threads
-                let mut local_reg = Self::new_dummy_with_constants(
+                let mut local_reg = Self::new_dummy_with_linear_coord(
                     starting_offset,
                     starting_linear_coord,
-                    starting_constants,
                     self.num_inner_cols,
-                    self.throw_range_check_error,
+                    self.witness_gen,
                 );
                 let res = inner_loop_function(idx, &mut local_reg);
                 // we update the offset and constants
@@ -343,10 +372,6 @@ impl<'a, F: PrimeField + TensorType + PartialOrd> RegionCtx<'a, F> {
                     local_reg.linear_coord() - starting_linear_coord,
                     Ordering::SeqCst,
                 );
-                constants.fetch_add(
-                    local_reg.total_constants() - starting_constants,
-                    Ordering::SeqCst,
-                );
 
                 max_lookup_inputs.fetch_max(local_reg.max_lookup_inputs(), Ordering::SeqCst);
                 min_lookup_inputs.fetch_min(local_reg.min_lookup_inputs(), Ordering::SeqCst);
@@ -362,11 +387,13 @@ impl<'a, F: PrimeField + TensorType + PartialOrd> RegionCtx<'a, F> {
                 // update the shuffle index
                 let mut shuffle_index = shuffle_index.lock().unwrap();
                 shuffle_index.update(&local_reg.shuffle_index);
+                // update the constants
+                let mut constants = constants.lock().unwrap();
+                constants.extend(local_reg.assigned_constants);
 
                 res
             })
             .map_err(|e| RegionError::from(format!("dummy_loop: {:?}", e)))?;
-        self.total_constants = constants.into_inner();
         self.linear_coord = linear_coord.into_inner();
         #[allow(trivial_numeric_casts)]
         {
@@ -410,6 +437,14 @@ impl<'a, F: PrimeField + TensorType + PartialOrd> RegionCtx<'a, F> {
             .map_err(|e| {
                 RegionError::from(format!("dummy_loop: failed to get shuffle index: {:?}", e))
             })?;
+        self.assigned_constants = Arc::try_unwrap(constants)
+            .map_err(|e| {
+                RegionError::from(format!("dummy_loop: failed to get constants: {:?}", e))
+            })?
+            .into_inner()
+            .map_err(|e| {
+                RegionError::from(format!("dummy_loop: failed to get constants: {:?}", e))
+            })?;
 
         Ok(())
     }
@@ -435,7 +470,7 @@ impl<'a, F: PrimeField + TensorType + PartialOrd> RegionCtx<'a, F> {
         range: Range,
     ) -> Result<(), Box<dyn std::error::Error>> {
         if range.0 > range.1 {
-            return Err("update_max_min_lookup_range: invalid range".into());
+            return Err(format!("update_max_min_lookup_range: invalid range {:?}", range).into());
         }
 
         let range_size = (range.1 - range.0).abs();
@@ -477,7 +512,7 @@ impl<'a, F: PrimeField + TensorType + PartialOrd> RegionCtx<'a, F> {
 
     /// Get the total number of constants
     pub fn total_constants(&self) -> usize {
-        self.total_constants
+        self.assigned_constants.len()
     }
 
     /// Get the dynamic lookup index
@@ -525,26 +560,24 @@ impl<'a, F: PrimeField + TensorType + PartialOrd> RegionCtx<'a, F> {
         self.max_range_size
     }
 
-    /// Assign a constant value
-    pub fn assign_constant(&mut self, var: &VarTensor, value: F) -> Result<ValType<F>, Error> {
-        self.total_constants += 1;
-        if let Some(region) = &self.region {
-            let cell = var.assign_constant(&mut region.borrow_mut(), self.linear_coord, value)?;
-            Ok(cell.into())
-        } else {
-            Ok(value.into())
-        }
-    }
     /// Assign a valtensor to a vartensor
     pub fn assign(
         &mut self,
         var: &VarTensor,
         values: &ValTensor<F>,
     ) -> Result<ValTensor<F>, Error> {
-        self.total_constants += values.num_constants();
         if let Some(region) = &self.region {
-            var.assign(&mut region.borrow_mut(), self.linear_coord, values)
+            var.assign(
+                &mut region.borrow_mut(),
+                self.linear_coord,
+                values,
+                &mut self.assigned_constants,
+            )
         } else {
+            if !values.is_instance() {
+                let values_map = values.create_constants_map_iterator();
+                self.assigned_constants.extend(values_map);
+            }
             Ok(values.clone())
         }
     }
@@ -560,14 +593,18 @@ impl<'a, F: PrimeField + TensorType + PartialOrd> RegionCtx<'a, F> {
         var: &VarTensor,
         values: &ValTensor<F>,
     ) -> Result<ValTensor<F>, Error> {
-        self.total_constants += values.num_constants();
         if let Some(region) = &self.region {
             var.assign(
                 &mut region.borrow_mut(),
                 self.combined_dynamic_shuffle_coord(),
                 values,
+                &mut self.assigned_constants,
             )
         } else {
+            if !values.is_instance() {
+                let values_map = values.create_constants_map_iterator();
+                self.assigned_constants.extend(values_map);
+            }
             Ok(values.clone())
         }
     }
@@ -594,13 +631,20 @@ impl<'a, F: PrimeField + TensorType + PartialOrd> RegionCtx<'a, F> {
                 self.linear_coord,
                 values,
                 ommissions,
+                &mut self.assigned_constants,
             )
         } else {
-            self.total_constants += values.num_constants();
             let inner_tensor = values.get_inner_tensor().unwrap();
+            let mut values_map = values.create_constants_map();
+
             for o in ommissions {
-                self.total_constants -= inner_tensor.get_flat_index(**o).is_constant() as usize;
+                if let ValType::Constant(value) = inner_tensor.get_flat_index(**o) {
+                    values_map.remove(&value);
+                }
             }
+
+            self.assigned_constants.extend(values_map);
+
             Ok(values.clone())
         }
     }
@@ -615,24 +659,24 @@ impl<'a, F: PrimeField + TensorType + PartialOrd> RegionCtx<'a, F> {
     ) -> Result<(ValTensor<F>, usize), Error> {
         if let Some(region) = &self.region {
             // duplicates every nth element to adjust for column overflow
-            let (res, len, total_assigned_constants) = var.assign_with_duplication(
+            let (res, len) = var.assign_with_duplication(
                 &mut region.borrow_mut(),
                 self.row,
                 self.linear_coord,
                 values,
                 check_mode,
                 single_inner_col,
+                &mut self.assigned_constants,
             )?;
-            self.total_constants += total_assigned_constants;
             Ok((res, len))
         } else {
-            let (_, len, total_assigned_constants) = var.dummy_assign_with_duplication(
+            let (_, len) = var.dummy_assign_with_duplication(
                 self.row,
                 self.linear_coord,
                 values,
                 single_inner_col,
+                &mut self.assigned_constants,
             )?;
-            self.total_constants += total_assigned_constants;
             Ok((values.clone(), len))
         }
     }
@@ -699,9 +743,4 @@ impl<'a, F: PrimeField + TensorType + PartialOrd> RegionCtx<'a, F> {
         }
         Ok(())
     }
-
-    /// increment constants
-    pub fn increment_constants(&mut self, n: usize) {
-        self.total_constants += n
-    }
 }