Skip to content

Support layerwise quantization #5201

Support layerwise quantization

Support layerwise quantization #5201

Workflow file for this run

name: OpenVINO - Test
on:
push:
branches:
- main
- v*-release
pull_request:
branches:
- main
concurrency:
group: ${{ github.workflow }}-${{ github.head_ref || github.run_id }}
cancel-in-progress: true
jobs:
build:
strategy:
fail-fast: false
matrix:
test-pattern:
[
"*modeling*",
"*diffusion*",
"*quantization*",
"*training*",
"*export*",
]
transformers-version: ["4.36.0", "latest"]
runs-on: ubuntu-22.04
steps:
- name: Checkout code
uses: actions/checkout@v4
- name: Setup Python
uses: actions/setup-python@v5
with:
python-version: 3.9
- name: Install dependencies
run: |
pip install --upgrade pip
pip install torch torchvision torchaudio --index-url https://download.pytorch.org/whl/cpu
pip install .[openvino,openvino-tokenizers,diffusers,tests] transformers[testing]
- if: ${{ matrix.transformers-version != 'latest' }}
name: Downgrade Transformers and Accelerate
run: |
pip install transformers==${{ matrix.transformers-version }} accelerate==0.*
- if: ${{ matrix.test-pattern == '*modeling*' }}
name: Uninstall NNCF
run: |
pip uninstall -y nncf
- name: Test with Pytest
run: |
pytest tests/openvino/${{ matrix.test-pattern }} --durations=0
env:
HF_HUB_READ_TOKEN: ${{ secrets.HF_HUB_READ_TOKEN }}
- if: ${{ matrix.test-pattern == '*modeling*' }}
name: Install dependencies (nightly)
run: |
pip install --upgrade --pre openvino openvino-tokenizers --extra-index-url https://storage.openvinotoolkit.org/simple/wheels/nightly
- if: ${{ matrix.test-pattern == '*modeling*' }}
name: Test with Pytest (nightly)
run: |
pytest tests/openvino/test_modeling_basic.py --durations=0