| name: NeuralOps CI & Reproducibility Pipeline | |
| on: | |
| push: | |
| branches: [ main ] | |
| pull_request: | |
| branches: [ main ] | |
| jobs: | |
| smoke-test-and-reproducibility: | |
| name: NeurIPS/ICLR CI Reproducibility Check | |
| runs-on: ubuntu-24.04 | |
| steps: | |
| - name: Checkout Codebase | |
| uses: actions/checkout@v7 | |
| - name: Set up Python 3.12 | |
| uses: actions/setup-python@v7 | |
| with: | |
| python-version: "3.12" | |
| - name: Install CPU-only PyTorch and Pinned Dependencies | |
| run: | | |
| python -m pip install --upgrade pip setuptools wheel | |
| pip install torch torchvision torchaudio --index-url https://download.pytorch.org/whl/cpu | |
| pip install transformers==4.46.3 datasets==3.1.0 scikit-learn==1.5.2 scipy==1.14.1 pandas==2.2.3 numpy==1.26.4 tqdm==4.67.1 matplotlib==3.9.2 PyYAML==6.0.2 pyarrow==18.0.0 codecarbon==2.7.2 fvcore==0.1.5.post20221221 accelerate==1.1.1 tabulate==0.9.0 pytest==8.3.3 | |
| - name: Syntax and Bytecode Check | |
| run: | | |
| python -m py_compile Codes/framework/attentions/*.py | |
| python -m py_compile Codes/framework/compression_engine/*.py | |
| python -m py_compile Codes/framework/*.py | |
| - name: Execute Minimal Benchmark Sweep (Smoke Test) | |
| env: | |
| PYTHONPATH: Codes:. | |
| run: | | |
| python Codes/framework/run_quality_sweep.py \ | |
| --finetune_steps 1 \ | |
| --max_steps 1 \ | |
| --seed 42 \ | |
| --modules full,a3,minimal_svd \ | |
| --model bert-base-uncased \ | |
| --model_family bert \ | |
| --dataset sst2 \ | |
| --batch_size 2 \ | |
| --seq_len 128 \ | |
| --output_dir /tmp/test | |
| test -f /tmp/test/summary.csv | |
| test -f /tmp/test/detailed_results.json | |
| cat /tmp/test/summary.csv | |
| - name: Run Core Pytest Suite | |
| env: | |
| PYTHONPATH: Codes:. | |
| run: | | |
| # Run a comprehensive subset of the test suite that can complete quickly on CI | |
| pytest Codes/framework/tests/test_attention_modules.py Codes/framework/tests/test_compressors.py Codes/framework/tests/test_sweep_orchestration.py Codes/framework/tests/test_lora.py -v --maxfail=5 | |
Xet Storage Details
- Size:
- 2.13 kB
- Xet hash:
- c9abcb3fddad58ab816d56727723ba4f30902f91db6efb6cadbdc27c67e90b84
·
Xet efficiently stores files, intelligently splitting them into unique chunks and accelerating uploads and downloads. More info.