LLAMA_CPP plugin - basic version with direct file loading #22
Workflow file for this run
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| name: llama_cpp_plugin_build_and_test | |
| on: | |
| pull_request: | |
| types: | |
| - opened | |
| - reopened | |
| - synchronize | |
| paths: | |
| - 'modules/llama_cpp_plugin/**' | |
| jobs: | |
| build_ubuntu20: | |
| runs-on: ubuntu-20.04 | |
| steps: | |
| - name: Setup cmake | |
| uses: jwlawson/actions-setup-cmake@v1.14 | |
| with: | |
| cmake-version: '3.24.x' | |
| - name: Checkout openvino_contrib | |
| uses: actions/checkout@v3 | |
| with: | |
| submodules: recursive | |
| path: openvino_contrib | |
| - name: Checkout openvino | |
| uses: actions/checkout@v3 | |
| with: | |
| submodules: recursive | |
| repository: vshampor/openvino | |
| ref: llama_cpp_mod | |
| path: openvino | |
| - name: CMake - configure | |
| run: cmake -B build -DCMAKE_BUILD_TYPE=Release -DOPENVINO_EXTRA_MODULES=${{ github.workspace }}/openvino_contrib/modules -DBUILD_java_api=OFF -DBUILD_nvidia_plugin=OFF -DBUILD_custom_operations=OFF -DBUILD_openvino_code=OFF -DBUILD_token_merging=OFF -DENABLE_TESTS=ON -DENABLE_FUNCTIONAL_TESTS=ON -DENABLE_PLUGINS_XML=ON openvino | |
| - name: CMake - build | |
| run: cmake --build build -j`nproc` -- llama_cpp_plugin llama_cpp_e2e_tests | |
| - name: Upload build artifacts | |
| uses: actions/upload-artifact@v4 | |
| with: | |
| name: build_artifacts | |
| path: ${{ github.workspace }}/openvino/bin/intel64/Release/ | |
| test_ubuntu20: | |
| needs: build_ubuntu20 | |
| runs-on: ubuntu-20.04 | |
| steps: | |
| - name: Download build artifacts | |
| uses: actions/download-artifact@v4 | |
| with: | |
| name: build_artifacts | |
| path: ${{ github.workspace }}/binaries | |
| - name: Prepare test data - checkout llama.cpp repo | |
| uses: actions/checkout@v3 | |
| with: | |
| repository: ggerganov/llama.cpp | |
| path: llama.cpp | |
| - name: Prepare test data - convert test model files | |
| run: | | |
| pip install -r llama.cpp/requirements/requirements-convert-hf-to-gguf.txt | |
| huggingface-cli download gpt2 model.safetensors tokenizer.json tokenizer_config.json vocab.json config.json merges.txt --local-dir hf_gpt2 | |
| mkdir -p ${{ github.workspace }}/test_data | |
| python3 llama.cpp/convert-hf-to-gguf.py hf_gpt2 --outtype f32 --outfile ${{ github.workspace }}/test_data/gpt2.gguf | |
| - name: Run E2E tests | |
| run: | | |
| chmod +x ${{ github.workspace }}/binaries/llama_cpp_e2e_tests | |
| ${{ github.workspace }}/binaries/llama_cpp_e2e_tests |