Weights & Biases makes running collaborative machine learning projects a breeze. You can focus on what you're trying to experiment with, and W&B will take on the burden of keeping track of everything. If you want to review a loss plot, download the latest model for production, or just see which configurations produced a certain model, W&B is your friend. There's also a bunch of features to help you and your team collaborate, like having a shared dashboard and sharing interactive reports.

How Weights and Biases can help you with Audits and Regulatory Guidelines

This notebook accompanies and implements a blog post on using W&B Artifacts to help teams in regulation-heavy industries share their Machine Learning models with clients.

Run the cells below to train an image classifier and upload the model checkpoints as W&B Artifacts. Then you can reliably know which models you've given to your clients and happily share this information with any regulators.

Please make sure that you set CUDA device before running the following colab. This can be done by changing Runtime Type to use GPU hardware accelerator.

Setup

✍️ Login to W&B

🏋️‍♀️ Model Training and Evaluation

🏁 Checkpoint Saver

Track top-n training checkpoints and maintain recovery checkpoints on specified intervals. Hacked together by / Copyright 2020 Ross Wightman

This script has been adapted from pytorch-image-models checkpoint saver script written by Ross Wightman. This script adds Weights and Biases artifact integration on top. (https://github.com/rwightman/pytorch-image-models/blob/master/timm/utils/checkpoint_saver.py)

## 💡Bring it all together!

Train Model and Log Artifacts to W&B

Upload Artifact to S3