MLproject

name: cifar10_initial

entry_points:
  main:
    parameters:
      preprocess_data: {type: string, default: cifar10}
      preprocess_downstream: {type: string, default: /opt/data/preprocess/}
      preprocess_cached_data_id: {type: string, default: ""}

      train_downstream: {type: string, default: /opt/data/model/}
      train_tensorboard: {type: string, default: /opt/data/tensorboard/}
      train_epochs: {type: int, default: 1}
      train_batch_size: {type: int, default: 32}
      train_num_workers: {type: int, default: 4}
      train_learning_rate: {type: float, default: 0.001}
      train_model_type: {type: string, default: vgg11}

      building_dockerfile_path: {type: string, default: ./Dockerfile}
      building_model_filename: {type: string, default: cifar10_0.onnx}
      building_entrypoint_path: {type: string, default: ./onnx_runtime_server_entrypoint.sh}

      evaluate_downstream: {type: string, default: ./evaluate/}
    command: |
      python -m main \
        --preprocess_data {preprocess_data} \
        --preprocess_downstream {preprocess_downstream} \
        --preprocess_cached_data_id {preprocess_cached_data_id} \
        --train_downstream {train_downstream} \
        --train_tensorboard {train_tensorboard} \
        --train_epochs {train_epochs} \
        --train_batch_size {train_batch_size} \
        --train_num_workers {train_num_workers} \
        --train_learning_rate {train_learning_rate} \
        --train_model_type {train_model_type} \
        --building_dockerfile_path {building_dockerfile_path} \
        --building_model_filename {building_model_filename} \
        --building_entrypoint_path {building_entrypoint_path} \
        --evaluate_downstream {evaluate_downstream}