mirror of
https://github.com/tinygrad/tinygrad.git
synced 2026-01-10 07:28:15 -05:00
ci set PYTHONPATH for all (#11997)
This commit is contained in:
chenyu
129
.github/workflows/test.yml
vendored
129
.github/workflows/test.yml
vendored
@@ -7,6 +7,7 @@ env:
|
||||
BUILD_CACHE_VERSION: '1'
|
||||
CAPTURE_PROCESS_REPLAY: 1
|
||||
GH_TOKEN: ${{ secrets.GITHUB_TOKEN }}
|
||||
PYTHONPATH: ${{ github.workspace }}
|
||||
|
||||
on:
|
||||
push:
|
||||
@@ -79,7 +80,7 @@ jobs:
|
||||
python docs/abstractions2.py
|
||||
python docs/abstractions3.py
|
||||
- name: Test Quickstart
|
||||
run: awk '/```python/{flag=1;next}/```/{flag=0}flag' docs/quickstart.md > quickstart.py && PYTHONPATH=. python quickstart.py
|
||||
run: awk '/```python/{flag=1;next}/```/{flag=0}flag' docs/quickstart.md > quickstart.py && python quickstart.py
|
||||
- name: Test DEBUG
|
||||
run: DEBUG=100 python3 -c "from tinygrad import Tensor; N = 1024; a, b = Tensor.rand(N, N), Tensor.rand(N, N); c = (a.reshape(N, 1, N) * b.T.reshape(1, N, N)).sum(axis=2); print((c.numpy() - (a.numpy() @ b.numpy())).mean())"
|
||||
- name: Compile EfficientNet to C and test it
|
||||
@@ -182,19 +183,19 @@ jobs:
|
||||
pip3 install --upgrade --force-reinstall ruff==0.11.0
|
||||
python3 -m ruff check extra/torch_backend/backend.py
|
||||
- name: Test one op
|
||||
run: PYTHONPATH=. FORWARD_ONLY=1 TINY_BACKEND=1 python3 test/test_ops.py TestOps.test_add
|
||||
run: FORWARD_ONLY=1 TINY_BACKEND=1 python3 test/test_ops.py TestOps.test_add
|
||||
- name: Test ResNet-18
|
||||
run: PYTHONPATH=. DEBUG=2 python3 extra/torch_backend/example.py
|
||||
run: DEBUG=2 python3 extra/torch_backend/example.py
|
||||
- name: My (custom) tests
|
||||
run: PYTHONPATH=. python3 extra/torch_backend/test.py
|
||||
run: python3 extra/torch_backend/test.py
|
||||
- name: Test one op in torch tests
|
||||
run: PYTHONPATH=. DEBUG=2 python3 extra/torch_backend/torch_tests.py TestTinyBackendPRIVATEUSE1.test_unary_log_tiny_float32
|
||||
run: DEBUG=2 python3 extra/torch_backend/torch_tests.py TestTinyBackendPRIVATEUSE1.test_unary_log_tiny_float32
|
||||
- name: Test Ops with TINY_BACKEND
|
||||
run: PYTHONPATH=. LLVM=1 LLVMOPT=0 TINY_BACKEND=1 python3 -m pytest -n auto test/test_ops.py --durations=20
|
||||
run: LLVM=1 LLVMOPT=0 TINY_BACKEND=1 python3 -m pytest -n auto test/test_ops.py --durations=20
|
||||
- name: Test in-place operations on views
|
||||
run: PYTHONPATH=. TORCH_DEBUG=1 python3 extra/torch_backend/test_inplace.py
|
||||
run: TORCH_DEBUG=1 python3 extra/torch_backend/test_inplace.py
|
||||
- name: Test multi-gpu
|
||||
run: PYTHONPATH=. LLVM=1 GPUS=4 TORCH_DEBUG=1 python3 extra/torch_backend/test_multigpu.py
|
||||
run: LLVM=1 GPUS=4 TORCH_DEBUG=1 python3 extra/torch_backend/test_multigpu.py
|
||||
|
||||
torchbackendmore:
|
||||
name: Torch Backend Tests More
|
||||
@@ -216,9 +217,9 @@ jobs:
|
||||
sudo apt update || true
|
||||
sudo apt install -y --no-install-recommends ninja-build
|
||||
- name: Test beautiful_mnist in torch with TINY_BACKEND
|
||||
run: SPLIT_REDUCEOP=0 FUSE_ARANGE=1 PYTHONPATH=. LLVM=1 TARGET_EVAL_ACC_PCT=96.0 TINY_BACKEND=1 python3 examples/other_mnist/beautiful_mnist_torch.py
|
||||
run: SPLIT_REDUCEOP=0 FUSE_ARANGE=1 LLVM=1 TARGET_EVAL_ACC_PCT=96.0 TINY_BACKEND=1 python3 examples/other_mnist/beautiful_mnist_torch.py
|
||||
- name: Test some torch tests (expect failure)
|
||||
run: PYTHONPATH=. python3 -m pytest extra/torch_backend/torch_tests.py -v --tb=no || true
|
||||
run: python3 -m pytest extra/torch_backend/torch_tests.py -v --tb=no || true
|
||||
|
||||
tc:
|
||||
name: Tensor Core tests
|
||||
@@ -241,31 +242,31 @@ jobs:
|
||||
- name: Test emulated METAL tensor cores
|
||||
run: |
|
||||
DEBUG=2 EMULATE_METAL=1 FORWARD_ONLY=1 PYTHON=1 python3 test/test_ops.py TestOps.test_big_gemm
|
||||
PYTHONPATH=. DEBUG=2 EMULATE_METAL=1 FORWARD_ONLY=1 PYTHON=1 python3 test/test_linearizer.py TestLinearizer.test_tensor_cores
|
||||
PYTHONPATH=. DEBUG=2 EMULATE_METAL=1 FORWARD_ONLY=1 PYTHON=1 python3 test/test_linearizer.py TestLinearizer.test_tensor_cores_padded TestLinearizer.test_tensor_cores_padded_uops
|
||||
DEBUG=2 EMULATE_METAL=1 FORWARD_ONLY=1 PYTHON=1 python3 test/test_linearizer.py TestLinearizer.test_tensor_cores
|
||||
DEBUG=2 EMULATE_METAL=1 FORWARD_ONLY=1 PYTHON=1 python3 test/test_linearizer.py TestLinearizer.test_tensor_cores_padded TestLinearizer.test_tensor_cores_padded_uops
|
||||
- name: Test emulated AMX tensor cores
|
||||
run: PYTHONPATH=. DEBUG=2 AMX=1 EMULATE_AMX=1 FORWARD_ONLY=1 PYTHON=1 python3 test/test_ops.py TestOps.test_gemm
|
||||
run: DEBUG=2 AMX=1 EMULATE_AMX=1 FORWARD_ONLY=1 PYTHON=1 python3 test/test_ops.py TestOps.test_gemm
|
||||
- name: Test emulated AMD tensor cores
|
||||
run: |
|
||||
PYTHONPATH=. DEBUG=2 EMULATE_AMD=1 FORWARD_ONLY=1 PYTHON=1 N=16 HALF=1 ACC_HALF=0 python3 ./extra/gemm/simple_matmul.py
|
||||
PYTHONPATH=. DEBUG=2 EMULATE_AMD=1 FORWARD_ONLY=1 PYTHON=1 N=64 HALF=1 ACC_HALF=0 python3 ./extra/gemm/simple_matmul.py
|
||||
PYTHONPATH=. DEBUG=2 EMULATE_AMD=1 FORWARD_ONLY=1 PYTHON=1 N=16 HALF=1 ACC_HALF=1 ATOL=1e-3 python3 ./extra/gemm/simple_matmul.py
|
||||
PYTHONPATH=. DEBUG=2 EMULATE_AMD=1 FORWARD_ONLY=1 PYTHON=1 N=64 HALF=1 ACC_HALF=1 ATOL=1e-3 python3 ./extra/gemm/simple_matmul.py
|
||||
PYTHONPATH=. DEBUG=2 EMULATE_AMD=1 FORWARD_ONLY=1 PYTHON=1 python3 test/test_linearizer.py TestLinearizer.test_tensor_cores
|
||||
PYTHONPATH=. DEBUG=2 EMULATE_AMD=1 FORWARD_ONLY=1 PYTHON=1 python3 test/test_linearizer.py TestLinearizer.test_tensor_cores_padded_amd TestLinearizer.test_tensor_cores_padded_uops
|
||||
DEBUG=2 EMULATE_AMD=1 FORWARD_ONLY=1 PYTHON=1 N=16 HALF=1 ACC_HALF=0 python3 ./extra/gemm/simple_matmul.py
|
||||
DEBUG=2 EMULATE_AMD=1 FORWARD_ONLY=1 PYTHON=1 N=64 HALF=1 ACC_HALF=0 python3 ./extra/gemm/simple_matmul.py
|
||||
DEBUG=2 EMULATE_AMD=1 FORWARD_ONLY=1 PYTHON=1 N=16 HALF=1 ACC_HALF=1 ATOL=1e-3 python3 ./extra/gemm/simple_matmul.py
|
||||
DEBUG=2 EMULATE_AMD=1 FORWARD_ONLY=1 PYTHON=1 N=64 HALF=1 ACC_HALF=1 ATOL=1e-3 python3 ./extra/gemm/simple_matmul.py
|
||||
DEBUG=2 EMULATE_AMD=1 FORWARD_ONLY=1 PYTHON=1 python3 test/test_linearizer.py TestLinearizer.test_tensor_cores
|
||||
DEBUG=2 EMULATE_AMD=1 FORWARD_ONLY=1 PYTHON=1 python3 test/test_linearizer.py TestLinearizer.test_tensor_cores_padded_amd TestLinearizer.test_tensor_cores_padded_uops
|
||||
- name: Test emulated AMD MFMA tensor cores
|
||||
run: |
|
||||
PYTHONPATH=. DEBUG=2 EMULATE_AMD_MFMA=1 FORWARD_ONLY=1 PYTHON=1 N=64 HALF=1 ACC_HALF=0 python3 ./extra/gemm/simple_matmul.py
|
||||
PYTHONPATH=. DEBUG=2 EMULATE_AMD_MFMA=1 FORWARD_ONLY=1 PYTHON=1 python3 test/test_linearizer.py TestLinearizer.test_tensor_cores
|
||||
PYTHONPATH=. DEBUG=2 EMULATE_AMD_MFMA=1 FORWARD_ONLY=1 PYTHON=1 python3 test/test_linearizer.py TestLinearizer.test_tensor_cores_padded TestLinearizer.test_tensor_cores_padded_uops
|
||||
DEBUG=2 EMULATE_AMD_MFMA=1 FORWARD_ONLY=1 PYTHON=1 N=64 HALF=1 ACC_HALF=0 python3 ./extra/gemm/simple_matmul.py
|
||||
DEBUG=2 EMULATE_AMD_MFMA=1 FORWARD_ONLY=1 PYTHON=1 python3 test/test_linearizer.py TestLinearizer.test_tensor_cores
|
||||
DEBUG=2 EMULATE_AMD_MFMA=1 FORWARD_ONLY=1 PYTHON=1 python3 test/test_linearizer.py TestLinearizer.test_tensor_cores_padded TestLinearizer.test_tensor_cores_padded_uops
|
||||
- name: Test emulated AMD RDNA4 tensor cores
|
||||
run: |
|
||||
PYTHONPATH=. DEBUG=2 EMULATE_AMD_RDNA4=1 FORWARD_ONLY=1 PYTHON=1 N=16 HALF=1 ACC_HALF=0 python3 ./extra/gemm/simple_matmul.py
|
||||
PYTHONPATH=. DEBUG=2 EMULATE_AMD_RDNA4=1 FORWARD_ONLY=1 PYTHON=1 N=64 HALF=1 ACC_HALF=0 python3 ./extra/gemm/simple_matmul.py
|
||||
PYTHONPATH=. DEBUG=2 EMULATE_AMD_RDNA4=1 FORWARD_ONLY=1 PYTHON=1 N=16 HALF=1 ACC_HALF=1 ATOL=1e-3 python3 ./extra/gemm/simple_matmul.py
|
||||
PYTHONPATH=. DEBUG=2 EMULATE_AMD_RDNA4=1 FORWARD_ONLY=1 PYTHON=1 N=64 HALF=1 ACC_HALF=1 ATOL=1e-3 python3 ./extra/gemm/simple_matmul.py
|
||||
PYTHONPATH=. DEBUG=2 EMULATE_AMD_RDNA4=1 FORWARD_ONLY=1 PYTHON=1 python3 test/test_linearizer.py TestLinearizer.test_tensor_cores
|
||||
PYTHONPATH=. DEBUG=2 EMULATE_AMD_RDNA4=1 FORWARD_ONLY=1 PYTHON=1 python3 test/test_linearizer.py TestLinearizer.test_tensor_cores_padded TestLinearizer.test_tensor_cores_padded_uops
|
||||
DEBUG=2 EMULATE_AMD_RDNA4=1 FORWARD_ONLY=1 PYTHON=1 N=16 HALF=1 ACC_HALF=0 python3 ./extra/gemm/simple_matmul.py
|
||||
DEBUG=2 EMULATE_AMD_RDNA4=1 FORWARD_ONLY=1 PYTHON=1 N=64 HALF=1 ACC_HALF=0 python3 ./extra/gemm/simple_matmul.py
|
||||
DEBUG=2 EMULATE_AMD_RDNA4=1 FORWARD_ONLY=1 PYTHON=1 N=16 HALF=1 ACC_HALF=1 ATOL=1e-3 python3 ./extra/gemm/simple_matmul.py
|
||||
DEBUG=2 EMULATE_AMD_RDNA4=1 FORWARD_ONLY=1 PYTHON=1 N=64 HALF=1 ACC_HALF=1 ATOL=1e-3 python3 ./extra/gemm/simple_matmul.py
|
||||
DEBUG=2 EMULATE_AMD_RDNA4=1 FORWARD_ONLY=1 PYTHON=1 python3 test/test_linearizer.py TestLinearizer.test_tensor_cores
|
||||
DEBUG=2 EMULATE_AMD_RDNA4=1 FORWARD_ONLY=1 PYTHON=1 python3 test/test_linearizer.py TestLinearizer.test_tensor_cores_padded TestLinearizer.test_tensor_cores_padded_uops
|
||||
- name: Test emulated CUDA tensor cores
|
||||
run: |
|
||||
DEBUG=2 EMULATE_CUDA=1 FORWARD_ONLY=1 PYTHON=1 python3 test/test_ops.py TestOps.test_gemm_fp16
|
||||
@@ -277,18 +278,18 @@ jobs:
|
||||
run: DEBUG=2 EMULATE_INTEL=1 FORWARD_ONLY=1 PYTHON=1 HALF=1 N=64 python3 ./extra/gemm/simple_matmul.py
|
||||
- name: Full test tensor cores
|
||||
run: |
|
||||
PYTHONPATH=. DEBUG=2 EMULATE_METAL=1 FORWARD_ONLY=1 PYTHON=1 python3 ./test/test_linearizer.py TestLinearizer.test_tensor_cores
|
||||
PYTHONPATH=. DEBUG=2 EMULATE_AMD=1 FORWARD_ONLY=1 PYTHON=1 python3 ./test/test_linearizer.py TestLinearizer.test_tensor_cores
|
||||
PYTHONPATH=. DEBUG=2 EMULATE_CUDA=1 ALLOW_TF32=1 FORWARD_ONLY=1 PYTHON=1 python3 ./test/test_linearizer.py TestLinearizer.test_tensor_cores
|
||||
PYTHONPATH=. DEBUG=2 EMULATE_INTEL=1 FORWARD_ONLY=1 PYTHON=1 python3 ./test/test_linearizer.py TestLinearizer.test_tensor_cores
|
||||
PYTHONPATH=. DEBUG=2 AMX=1 EMULATE_AMX=1 FORWARD_ONLY=1 PYTHON=1 python3 ./test/test_linearizer.py TestLinearizer.test_tensor_cores
|
||||
DEBUG=2 EMULATE_METAL=1 FORWARD_ONLY=1 PYTHON=1 python3 ./test/test_linearizer.py TestLinearizer.test_tensor_cores
|
||||
DEBUG=2 EMULATE_AMD=1 FORWARD_ONLY=1 PYTHON=1 python3 ./test/test_linearizer.py TestLinearizer.test_tensor_cores
|
||||
DEBUG=2 EMULATE_CUDA=1 ALLOW_TF32=1 FORWARD_ONLY=1 PYTHON=1 python3 ./test/test_linearizer.py TestLinearizer.test_tensor_cores
|
||||
DEBUG=2 EMULATE_INTEL=1 FORWARD_ONLY=1 PYTHON=1 python3 ./test/test_linearizer.py TestLinearizer.test_tensor_cores
|
||||
DEBUG=2 AMX=1 EMULATE_AMX=1 FORWARD_ONLY=1 PYTHON=1 python3 ./test/test_linearizer.py TestLinearizer.test_tensor_cores
|
||||
- name: Test device flop counts
|
||||
run: |
|
||||
PYTHONPATH=. DEBUG=2 EMULATE_METAL=1 PYTHON=1 python3 ./test/test_uops_stats.py TestUOpsStatsMatmulHalf
|
||||
PYTHONPATH=. DEBUG=2 EMULATE_AMD=1 PYTHON=1 python3 ./test/test_uops_stats.py TestUOpsStatsMatmulHalf
|
||||
PYTHONPATH=. DEBUG=2 EMULATE_CUDA=1 PYTHON=1 python3 ./test/test_uops_stats.py TestUOpsStatsMatmulHalf
|
||||
PYTHONPATH=. DEBUG=2 EMULATE_INTEL=1 PYTHON=1 python3 ./test/test_uops_stats.py TestUOpsStatsMatmulHalf
|
||||
PYTHONPATH=. DEBUG=2 AMX=1 EMULATE_AMX=1 PYTHON=1 python3 ./test/test_uops_stats.py TestUOpsStats.test_simple_matmul
|
||||
DEBUG=2 EMULATE_METAL=1 PYTHON=1 python3 ./test/test_uops_stats.py TestUOpsStatsMatmulHalf
|
||||
DEBUG=2 EMULATE_AMD=1 PYTHON=1 python3 ./test/test_uops_stats.py TestUOpsStatsMatmulHalf
|
||||
DEBUG=2 EMULATE_CUDA=1 PYTHON=1 python3 ./test/test_uops_stats.py TestUOpsStatsMatmulHalf
|
||||
DEBUG=2 EMULATE_INTEL=1 PYTHON=1 python3 ./test/test_uops_stats.py TestUOpsStatsMatmulHalf
|
||||
DEBUG=2 AMX=1 EMULATE_AMX=1 PYTHON=1 python3 ./test/test_uops_stats.py TestUOpsStats.test_simple_matmul
|
||||
|
||||
bepython:
|
||||
name: Python Backend
|
||||
@@ -305,15 +306,15 @@ jobs:
|
||||
key: be-minimal
|
||||
deps: testing_minimal
|
||||
- name: Test dtype with Python emulator
|
||||
run: DEBUG=1 PYTHONPATH=. PYTHON=1 python3 -m pytest -n=auto test/test_dtype.py test/test_dtype_alu.py
|
||||
run: DEBUG=1 PYTHON=1 python3 -m pytest -n=auto test/test_dtype.py test/test_dtype_alu.py
|
||||
- name: Test ops with Python emulator
|
||||
run: DEBUG=2 PYTHON=1 python3 -m pytest -n=auto test/test_ops.py -k "not (test_split or test_simple_cumsum or test_cumsum or test_einsum or test_dot or test_dot_1d or test_big_gemm or test_broadcastdot or test_multidot or test_var_axis or test_std_axis or test_broadcast_full or test_broadcast_partial or test_simple_conv3d or test_dilated_conv_transpose2d or test_simple_conv_transpose3d or test_large_input_conv2d or test_max_pool2d or test_max_pool2d_simple or test_max_pool2d_bigger_stride or test_avg_pool2d or test_cat or test_scaled_product_attention or test_scaled_product_attention_causal or test_slice_fancy_indexing_dim_inject_none or test_slice_fancy_indexing_list_indices or test_slice_fancy_indexing_no_dim_collapse or test_slice_fancy_indexing_tuple_indices or test_slice_fancy_indexing_list_with_tensors or test_slice_fancy_indexing_dim_collapse_int or test_interpolate_bilinear or test_interpolate_bilinear_corners_aligned or test_scaled_dot_product_attention or test_cummax or test_simple_cummax or test_logcumsumexp or test_sort or test_cumprod)" --durations=20
|
||||
- name: Test uops with Python emulator
|
||||
run: PYTHON=1 python3 -m pytest test/test_uops.py --durations=20
|
||||
- name: Test symbolic with Python emulator
|
||||
run: PYTHONPATH=. PYTHON=1 python3 test/test_symbolic_ops.py
|
||||
run: PYTHON=1 python3 test/test_symbolic_ops.py
|
||||
- name: test_renderer_failures with Python emulator
|
||||
run: PYTHONPATH=. PYTHON=1 python3 -m pytest -rA test/test_renderer_failures.py::TestRendererFailures
|
||||
run: PYTHON=1 python3 -m pytest -rA test/test_renderer_failures.py::TestRendererFailures
|
||||
|
||||
linter:
|
||||
name: Linters
|
||||
@@ -361,7 +362,7 @@ jobs:
|
||||
pydeps: "pillow"
|
||||
deps: testing_unit
|
||||
- name: Test README
|
||||
run: awk '/```python/{flag=1;next}/```/{flag=0}flag' README.md > README.py && PYTHONPATH=. python README.py
|
||||
run: awk '/```python/{flag=1;next}/```/{flag=0}flag' README.md > README.py && python README.py
|
||||
- name: Run unit tests
|
||||
run: PYTHONPATH="." python -m pytest -n=auto test/unit/ --durations=20
|
||||
- name: Run targetted tests on NULL backend
|
||||
@@ -379,9 +380,9 @@ jobs:
|
||||
run: |
|
||||
test/external/process_replay/reset.py
|
||||
CAPTURE_PROCESS_REPLAY=1 python test/test_tiny.py TestTiny.test_plus
|
||||
PYTHONPATH=. python extra/optimization/extract_dataset.py
|
||||
python extra/optimization/extract_dataset.py
|
||||
gzip -c /tmp/sops > extra/datasets/sops.gz
|
||||
DEBUG=1 MIN_ASTS=1 PYTHONPATH=. python extra/optimization/get_action_space.py
|
||||
DEBUG=1 MIN_ASTS=1 python extra/optimization/get_action_space.py
|
||||
- name: Repo line count < 18000 lines
|
||||
run: MAX_LINE_COUNT=18000 python sz.py
|
||||
|
||||
@@ -513,11 +514,11 @@ jobs:
|
||||
- name: Test ONNX (LLVM)
|
||||
run: LLVM=1 python -m pytest -n=auto test/external/external_test_onnx_backend.py --durations=20
|
||||
- name: Test ONNX Runner (CPU)
|
||||
run: CPU=1 PYTHONPATH=. python3 test/external/external_test_onnx_runner.py
|
||||
run: CPU=1 python3 test/external/external_test_onnx_runner.py
|
||||
- name: Test Additional ONNX Ops (CPU)
|
||||
run: CPU=1 PYTHONPATH=. python3 test/external/external_test_onnx_ops.py
|
||||
run: CPU=1 python3 test/external/external_test_onnx_ops.py
|
||||
- name: Test Quantize ONNX
|
||||
run: CPU=1 PYTHONPATH=. python3 test/test_quantize_onnx.py
|
||||
run: CPU=1 python3 test/test_quantize_onnx.py
|
||||
- name: Run process replay tests
|
||||
uses: ./.github/actions/process-replay
|
||||
|
||||
@@ -542,15 +543,15 @@ jobs:
|
||||
- name: Test ONNX (GPU)
|
||||
run: GPU=1 python -m pytest -n=auto test/external/external_test_onnx_backend.py --durations=20
|
||||
- name: Test Optimization Helpers
|
||||
run: PYTHONPATH="." DEBUG=1 python3 extra/optimization/test_helpers.py
|
||||
run: DEBUG=1 python3 extra/optimization/test_helpers.py
|
||||
#- name: Test Action Space
|
||||
# run: PYTHONPATH="." DEBUG=1 GPU=1 python3 extra/optimization/get_action_space.py
|
||||
# run: DEBUG=1 GPU=1 python3 extra/optimization/get_action_space.py
|
||||
- name: Test Beam Search
|
||||
run: PYTHONPATH="." GPU=1 IGNORE_BEAM_CACHE=1 python3 -m pytest extra/optimization/test_beam_search.py
|
||||
run: GPU=1 IGNORE_BEAM_CACHE=1 python3 -m pytest extra/optimization/test_beam_search.py
|
||||
- name: Test MLPerf stuff
|
||||
run: GPU=1 python -m pytest -n=auto test/external/external_test_optim.py test/external/external_test_losses.py test/external/external_test_metrics.py test/external/external_test_datasets.py --durations=20
|
||||
- name: Test llama 3 training
|
||||
run: MAX_BUFFER_SIZE=0 PYTHONPATH="." DEV=NULL SAMPLES=300 BS=8 SEQLEN=512 GRADIENT_ACC_STEPS=8 FAKEDATA=1 DEFAULT_FLOAT=bfloat16 OPTIM_DTYPE=bfloat16 LLAMA3_SIZE=1B MODEL=llama3 python3 examples/mlperf/model_train.py
|
||||
run: MAX_BUFFER_SIZE=0 DEV=NULL SAMPLES=300 BS=8 SEQLEN=512 GRADIENT_ACC_STEPS=8 FAKEDATA=1 DEFAULT_FLOAT=bfloat16 OPTIM_DTYPE=bfloat16 LLAMA3_SIZE=1B MODEL=llama3 python3 examples/mlperf/model_train.py
|
||||
- name: Run process replay tests
|
||||
uses: ./.github/actions/process-replay
|
||||
|
||||
@@ -639,7 +640,7 @@ jobs:
|
||||
- name: Test LLVM=1 DEVECTORIZE=0
|
||||
run: LLVM=1 DEVECTORIZE=0 python3 -m pytest -n auto test/test_tiny.py test/test_ops.py -k "not test_avg_pool3d_failure"
|
||||
- name: Test LLVM=1 DEVECTORIZE=0 for model
|
||||
run: PYTHONPATH="." LLVM=1 DEVECTORIZE=0 python3 test/models/test_efficientnet.py
|
||||
run: LLVM=1 DEVECTORIZE=0 python3 test/models/test_efficientnet.py
|
||||
- name: Test CPU=1 DEVECTORIZE=0
|
||||
run: CPU=1 DEVECTORIZE=0 FUSE_ARANGE=0 python3 -m pytest -n auto test/test_tiny.py test/test_ops.py -k "not test_avg_pool3d_failure"
|
||||
|
||||
@@ -675,9 +676,9 @@ jobs:
|
||||
- name: Run test_tiny on DSP
|
||||
run: DEBUG=2 DSP=1 python test/test_tiny.py
|
||||
- name: Test transcendentals
|
||||
run: CC=clang-20 PYTHONPATH="." DEBUG=2 DSP=1 python test/test_transcendental.py TestTranscendentalVectorized
|
||||
run: CC=clang-20 DEBUG=2 DSP=1 python test/test_transcendental.py TestTranscendentalVectorized
|
||||
- name: Test quantize onnx
|
||||
run: PYTHONPATH="." DEBUG=2 DSP=1 python3 test/test_quantize_onnx.py
|
||||
run: DEBUG=2 DSP=1 python3 test/test_quantize_onnx.py
|
||||
|
||||
testwebgpu:
|
||||
name: Linux (WebGPU)
|
||||
@@ -720,7 +721,6 @@ jobs:
|
||||
MOCKGPU: 1
|
||||
FORWARD_ONLY: 1
|
||||
AMD_LLVM: ${{ matrix.backend == 'amdllvm' && '1' || matrix.backend != 'amdllvm' && '0' }}
|
||||
PYTHONPATH: ${{ github.workspace }}
|
||||
steps:
|
||||
- name: Checkout Code
|
||||
uses: actions/checkout@v4
|
||||
@@ -777,8 +777,8 @@ jobs:
|
||||
run: printf "${{ matrix.backend == 'PTX' && 'CUDA=1\nPTX=1' || matrix.backend == 'nv' && 'NV=1' }}" >> $GITHUB_ENV
|
||||
- name: Check Device.DEFAULT and print some source
|
||||
run: |
|
||||
PYTHONPATH=${{ github.workspace }} python3 -c "from tinygrad import Device; assert Device.DEFAULT in ['CUDA','NV'], Device.DEFAULT"
|
||||
DEBUG=5 PYTHONPATH=${{ github.workspace }} FORWARD_ONLY=1 python3 test/test_ops.py TestOps.test_add
|
||||
python3 -c "from tinygrad import Device; assert Device.DEFAULT in ['CUDA','NV'], Device.DEFAULT"
|
||||
DEBUG=5 FORWARD_ONLY=1 python3 test/test_ops.py TestOps.test_add
|
||||
- name: Run pytest (cuda)
|
||||
# skip multitensor because it's slow
|
||||
run: python -m pytest -n=auto test/ --ignore=test/models --ignore=test/unit --ignore test/test_gc.py --ignore test/test_multitensor.py --durations=20
|
||||
@@ -811,8 +811,8 @@ jobs:
|
||||
run: printf "${{ matrix.backend == 'llvm' && 'LLVM=1' || matrix.backend == 'cpu' && 'CPU=1' || matrix.backend == 'gpu' && 'GPU=1' }}" >> $GITHUB_ENV
|
||||
- name: Check Device.DEFAULT and print some source
|
||||
run: |
|
||||
PYTHONPATH=${{ github.workspace }} python3 -c "from tinygrad import Device; assert Device.DEFAULT in ['LLVM','CPU','GPU'], Device.DEFAULT"
|
||||
DEBUG=5 PYTHONPATH=${{ github.workspace }} FORWARD_ONLY=1 python3 test/test_ops.py TestOps.test_add
|
||||
python3 -c "from tinygrad import Device; assert Device.DEFAULT in ['LLVM','CPU','GPU'], Device.DEFAULT"
|
||||
DEBUG=5 FORWARD_ONLY=1 python3 test/test_ops.py TestOps.test_add
|
||||
- name: Run pytest (not cuda)
|
||||
run: python -m pytest -n=auto test/ --ignore=test/models --ignore=test/unit --durations=20
|
||||
- name: Run TRANSCENDENTAL math
|
||||
@@ -853,11 +853,11 @@ jobs:
|
||||
- name: Test tensor core ops (real)
|
||||
run: METAL=1 DEBUG=3 python test/test_ops.py TestOps.test_big_gemm
|
||||
- name: Test LLaMA compile speed
|
||||
run: PYTHONPATH="." METAL=1 python test/external/external_test_speed_llama.py
|
||||
run: METAL=1 python test/external/external_test_speed_llama.py
|
||||
- name: Test Beam Search
|
||||
run: PYTHONPATH="." METAL=1 IGNORE_BEAM_CACHE=1 python3 -m pytest extra/optimization/test_beam_search.py
|
||||
run: METAL=1 IGNORE_BEAM_CACHE=1 python3 -m pytest extra/optimization/test_beam_search.py
|
||||
#- name: Fuzz Test linearizer
|
||||
# run: PYTHONPATH="." METAL=1 DEPTH=4 FUZZ_N=50 FUZZ_MAX_SIZE=1000000 python test/external/fuzz_linearizer.py
|
||||
# run: METAL=1 DEPTH=4 FUZZ_N=50 FUZZ_MAX_SIZE=1000000 python test/external/fuzz_linearizer.py
|
||||
- name: Run TRANSCENDENTAL math
|
||||
run: TRANSCENDENTAL=2 python -m pytest -n=auto test/test_ops.py::TestOps::test_sin test/test_ops.py::TestOps::test_cos test/test_ops.py::TestOps::test_tan test/test_ops.py::TestOps::test_exp test/test_ops.py::TestOps::test_log --durations=20
|
||||
- name: Run pytest (amd)
|
||||
@@ -920,7 +920,7 @@ jobs:
|
||||
# cp $GITHUB_WORKSPACE/test/web/test_viz.js .
|
||||
# node test_viz.js
|
||||
- name: Test ONNX Runner (WEBGPU)
|
||||
run: WEBGPU=1 PYTHONPATH=. python3 test/external/external_test_onnx_runner.py
|
||||
run: WEBGPU=1 python3 test/external/external_test_onnx_runner.py
|
||||
|
||||
osxremote:
|
||||
name: MacOS (remote metal)
|
||||
@@ -952,7 +952,6 @@ jobs:
|
||||
timeout-minutes: 20
|
||||
env:
|
||||
REMOTE: 1
|
||||
PYTHONPATH: ${{ github.workspace }}
|
||||
steps:
|
||||
- name: Checkout Code
|
||||
uses: actions/checkout@v4
|
||||
@@ -1073,7 +1072,7 @@ jobs:
|
||||
run: printf "${{ matrix.backend == 'llvm' && 'LLVM=1' || matrix.backend == 'cpu' && 'CPU=1' || matrix.backend == 'webgpu' && 'WEBGPU=1'}}" >> $GITHUB_ENV
|
||||
- name: Run unit tests
|
||||
if: matrix.backend=='llvm'
|
||||
run: python -m pytest -n=auto test/unit/ --ignore=test/unit/test_disk_tensor.py --ignore=test/unit/test_elf.py --ignore=test/unit/test_tar.py
|
||||
run: python -m pytest -n=auto test/unit/ --ignore=test/unit/test_disk_tensor.py --ignore=test/unit/test_elf.py --ignore=test/unit/test_tar.py --durations=20
|
||||
- name: Run pytest (${{ matrix.backend }})
|
||||
shell: bash
|
||||
run: |
|
||||
|
||||
Reference in New Issue
Block a user