train-model: Array Sorting @ EC2 Spot #29
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: GPT Array Sorter Experiment | |
| on: | |
| workflow_dispatch: | |
| inputs: | |
| environment: | |
| description: 'Loss Function' | |
| required: true | |
| default: 'CrossEntropyLoss' | |
| type: choice | |
| options: | |
| - 'CrossEntropyLoss' | |
| number_of_blocks: | |
| description: 'The number of transformer blocks.' | |
| required: true | |
| type: string | |
| default: '2' | |
| coordinates: | |
| description: 'The word embedding dimension.' | |
| required: true | |
| type: string | |
| default: '9' | |
| words: | |
| description: 'The number of tokens in the input sequence.' | |
| required: true | |
| type: string | |
| default: '11' | |
| tokens: | |
| description: 'The number of tokens in the vocab.' | |
| required: true | |
| type: string | |
| default: '3' | |
| number_of_batches: | |
| description: 'The number of batches.' | |
| required: true | |
| type: string | |
| default: '32' | |
| number_of_epochs: | |
| description: 'The number of epochs.' | |
| required: true | |
| type: string | |
| default: '20' | |
| learning_rate: | |
| description: 'The learning rate (fixed).' | |
| required: true | |
| type: string | |
| default: '0.001' | |
| ami_id: | |
| description: 'The ami id.' | |
| required: true | |
| type: string | |
| default: 'ami-093cb9fb2d34920ad' | |
| instance_type: | |
| description: 'The instance type.' | |
| required: true | |
| type: string | |
| default: 'c5n.xlarge' | |
| permissions: | |
| contents: read | |
| jobs: | |
| build: | |
| runs-on: ubuntu-latest | |
| steps: | |
| - uses: actions/checkout@v3 | |
| - name: Set up Python | |
| uses: actions/setup-python@v2 | |
| with: | |
| python-version: '3.10' | |
| - name: Install Dependencies | |
| run: | | |
| python -m pip install --upgrade pip | |
| pip install awscli boto3==1.34.8 pydantic==2.5.3 pydantic-settings==2.1.0 mlflow==2.9.2 | |
| - name: Print Inputs | |
| run: | | |
| echo "coordinates" | |
| echo ${{ github.event.inputs.coordinates }} | |
| echo "tokens" | |
| echo ${{ github.event.inputs.tokens }} | |
| echo "words" | |
| echo ${{ github.event.inputs.words }} | |
| echo "number_of_blocks" | |
| echo ${{ github.event.inputs.number_of_blocks }} | |
| echo "environment" | |
| echo ${{ github.event.inputs.environment }} | |
| echo "number_of_batches" | |
| echo ${{ github.event.inputs.number_of_batches }} | |
| echo "number_of_epochs" | |
| echo ${{ github.event.inputs.number_of_epochs }} | |
| - name: Babysit Experiment | |
| run: | | |
| export MLFLOW_TRACKING_URL=${{ secrets.MLFLOW_TRACKING_URL }} | |
| export MLFLOW_TRACKING_USERNAME=${{ secrets.MLFLOW_TRACKING_USERNAME }} | |
| export MLFLOW_TRACKING_PASSWORD=${{ secrets.MLFLOW_TRACKING_PASSWORD }} | |
| export MLFLOW_EXPERIMENT_ID=${{ secrets.MLFLOW_EXPERIMENT_ID }} | |
| export AWS_ACCESS_KEY_ID=${{ secrets.AWS_ACCESS_KEY_ID }} | |
| export AWS_SECRET_ACCESS_KEY=${{ secrets.AWS_SECRET_ACCESS_KEY }} | |
| export REGION_NAME=eu-west-2 | |
| export AMI_ID=${{ github.event.inputs.ami_id }} | |
| export INSTANCE_TYPE=${{ github.event.inputs.instance_type }} | |
| export COORDINATES=${{ github.event.inputs.coordinates }} | |
| export TOKENS=${{ github.event.inputs.tokens }} | |
| export WORDS=${{ github.event.inputs.words }} | |
| export NUMBER_OF_BLOCKS=${{ github.event.inputs.number_of_blocks }} | |
| export LOSS_FUNCTION=${{ github.event.inputs.environment }} | |
| export NUMBER_OF_BATCHES=${{ github.event.inputs.number_of_batches }} | |
| export NUMBER_OF_EPOCHS=${{ github.event.inputs.number_of_epochs }} | |
| export LEARNING_RATE=${{ github.event.inputs.learning_rate }} | |
| cd gpt_array_sorter | |
| python train_main.py |