Skip to content

Client

cli

def cli(config)

pipeline

def pipeline(config, **kwargs)

Runs all components for which run=True in config file.

All activated (run=True) components from config file will be run in sequence. Options overwrite the config file. The config file is the only way to define pipeline components.

Args
  • config-file: Central configuration file.

dataprep

def dataprep(config, **kwargs)

Run data preparation and create job dir.

Creates a directory (job_dir) with the following files:

  • train_samples.json

  • val_samples.json

  • test_samples.json

  • class_mapping.json

train

def train(config, **kwargs)

Train a CNN.

Fine-tunes an ImageNet pre-trained CNN. The number of classes are derived from train_samples.json. After each epoch the model will be evaluated on val_samples.json.

The best model (based on valuation accuracy) will be saved.

Args

  • image_dir: Directory with image files.

  • job_dir: Directory with train_samples, val_samples, and class_mapping.json.

evaluate

def evaluate(config, **kwargs)

Evaluate a trained model.

Evaluation will be performed on test_samples.json.

Args

  • image_dir: Directory with image files.

  • job_dir: Directory with test_samples.json and class_mapping.json.

cloud

def cloud(config, **kwargs)

Launch/destroy a cloud compute instance.

Launch/destroy cloud instances with Terraform based on Terraform files in tf_dir.