Update model.yml with specified quant #18
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: Update model.yml with specified quant | |
| on: | |
| workflow_dispatch: | |
| inputs: | |
| key_value_pairs: | |
| description: "the 'key=value' pairs that you want to update, separated by space" | |
| required: true | |
| type: string | |
| target_model_id: | |
| description: "Target HuggingFace model ID to update. For ex: llama3.1" | |
| required: true | |
| type: string | |
| source_model_size: | |
| description: "The model size. For ex: 8b" | |
| required: true | |
| type: string | |
| quantization_level: | |
| description: "Quantization level (e.g., 'q4-km') or 'all' for all levels" | |
| required: true | |
| type: string | |
| default: 'all' | |
| env: | |
| USER_NAME: cortexso | |
| KEY_VALUE_PAIRS: ${{ inputs.key_value_pairs }} | |
| SOURCE_MODEL_SIZE: ${{ inputs.source_model_size }} | |
| TARGET_MODEL_ID: ${{ inputs.target_model_id }} | |
| QUANT_LEVEL: ${{ inputs.quantization_level }} | |
| jobs: | |
| model-yaml-updater: | |
| runs-on: ubuntu-20-04-gguf | |
| timeout-minutes: 7200 | |
| steps: | |
| - name: Checkout | |
| uses: actions/checkout@v4 | |
| with: | |
| submodules: recursive | |
| - name: Set up Python | |
| uses: actions/setup-python@v5 | |
| with: | |
| python-version: '3.10' | |
| - name: Cache Python packages | |
| uses: actions/cache@v4 | |
| with: | |
| path: | | |
| ~/.cache/pip | |
| ~/.local/share/pip | |
| .venv | |
| key: ${{ runner.os }}-pip-${{ github.sha }} | |
| restore-keys: | | |
| ${{ runner.os }}-pip- | |
| - name: Install dependencies | |
| run: | | |
| pip3 install -I hf-transfer huggingface-hub ruamel.yaml | |
| git lfs install | |
| - name: Prepare folders | |
| run: | | |
| rm -rf /mnt/models/${{ env.MODEL_NAME }}/yaml/ | |
| mkdir -p /mnt/models/${{ env.MODEL_NAME }}/yaml | |
| - name: Quantize and Upload | |
| run: | | |
| if [ "${{ env.QUANT_LEVEL }}" = "all" ]; then | |
| quant_levels=("q2-k" "q3-ks" "q3-km" "q3-kl" "q4-ks" "q4-km" "q5-ks" "q5-km" "q6-k" "q8-0") | |
| else | |
| quant_levels=("${{ env.QUANT_LEVEL }}") | |
| fi | |
| for quant in "${quant_levels[@]}"; do | |
| mkdir -p "/mnt/models/${{ env.MODEL_NAME }}/yaml/${quant}/" | |
| python3 scripts/update_model_yml.py \ | |
| --repo_id "${{ env.USER_NAME }}/${{ env.TARGET_MODEL_ID }}" \ | |
| --filename model.yml \ | |
| --branch "${{ env.SOURCE_MODEL_SIZE }}-gguf-${quant}" \ | |
| --save_path "/mnt/models/${{ env.MODEL_NAME }}/yaml/${quant}/" \ | |
| --key_value_pairs name=${{ env.TARGET_MODEL_ID }}:${{ env.SOURCE_MODEL_SIZE }}-gguf-${quant} ${{ env.KEY_VALUE_PAIRS }} | |
| done | |
| - name: Upload to Hugging Face | |
| run: | | |
| huggingface-cli login --token ${{ secrets.HUGGINGFACE_TOKEN_WRITE }} --add-to-git-credential | |
| if [ "${{ env.QUANT_LEVEL }}" = "all" ]; then | |
| quant_levels=("q2-k" "q3-ks" "q3-km" "q3-kl" "q4-ks" "q4-km" "q5-ks" "q5-km" "q6-k" "q8-0") | |
| else | |
| quant_levels=("${{ env.QUANT_LEVEL }}") | |
| fi | |
| for quant in "${quant_levels[@]}"; do | |
| huggingface-cli upload "${{ env.USER_NAME }}/${{ env.TARGET_MODEL_ID }}" \ | |
| "/mnt/models/${{ env.MODEL_NAME }}/yaml/${quant}/" . \ | |
| --revision "${{ env.SOURCE_MODEL_SIZE }}-gguf-${quant}" | |
| done | |
| rm -rf /mnt/models/${{ env.MODEL_NAME }}/yaml/* | |
| huggingface-cli logout |