Flax is a neural network library for JAX that is designed for flexibility.

Composable transformations of Python+NumPy programs: differentiate, vectorize, JIT to GPU/TPU, and more

Easily turn large sets of image urls to an image dataset. Can download, resize and package 100M urls in 20h on one machine.

🤗 Transformers: State-of-the-art Natural Language Processing for TensorFlow 2.0 and PyTorch.

DataComp: In search of the next generation of multimodal datasets

Screenshots with JavaScript

JAX implementation ViT-VQGAN

JAX implementation of VQGAN

Emoji for everyone. https://twemoji.twitter.com/

FID computation in Jax/Flax.

DALLE-tools provided useful dataset utilities to improve you workflow with WebDatasets.

Tensors and Dynamic neural networks in Python with strong GPU acceleration

Gather modern English word frequencies from all enwiki articles.

Example deep learning projects that use wandb's features.

Google Research

🤗 The largest hub of ready-to-use NLP datasets for ML models with fast, easy-to-use and efficient data manipulation tools

Implementation / replication of DALL-E, OpenAI's Text to Image Transformer, in Pytorch

Easily compute clip embeddings and build a clip retrieval system with them

The lightweight PyTorch wrapper for ML researchers. Scale your models. Write less boilerplate