VESSL examples

This repository contains VESSL examples. If you want to learn more about VESSL please follow the quick start documentation.

Use hyperparameters on VESSL

Hyperparameters are automatically add to the container as environment variables with the given key and value. If you want to use them at runtime, then append them to the start command as follows.

# Add learning_rate as a hyperparameter
python main.py --learning-rate $learning_rate

You can now take the desired type of argument in Python script at runtime.

import argparse

parser = argparse.ArgumentParser()
parser.add_argument('--learning-rate', type=float, default=0.01)
args = parser.parse_args()

Try out VESSL examples

• Image classification (CIFAR) using Convnets (PyTorch)
• Object detection with balloon dataset using Detectron2 (PyTorch)
• Distributed training for image classification (CIFAR) with Resnet18
• Language modeling using LSTM RNNs (PyTorch)
• Image classification (MNIST) using Convnets (PyTorch, Keras)