Stars
NanoGPT (124M) quality in 7.8 8xH100-minutes
MoVQGAN - model for the image encoding and reconstruction
Implementation of the specific Transformer architecture from PaLM - Scaling Language Modeling with Pathways - in Jax (Equinox framework)
Flax is a neural network library for JAX that is designed for flexibility.
The AI Scientist: Towards Fully Automated Open-Ended Scientific Discovery π§βπ¬
JAX implementation of the Mistral 7b v0.2 model
DeepMind's Tacotron-2 Tensorflow implementation
A multi-voice TTS system trained with an emphasis on quality
Hybrid ML physics model of the Earth's atmosphere
Simple single-file baselines for Q-Learning in pure-GPU setting
Cost aware hyperparameter tuning algorithm
Simplifying reinforcement learning for complex game environments
π OpenHands: Code Less, Make More
Composable transformations of Python NumPy programs: differentiate, vectorize, JIT to GPU/TPU, and more
The release codes of LA-MCTS with its application to Neural Architecture Search.
Official codebase used to develop Vision Transformer, SigLIP, MLP-Mixer, LiT and more.
Train vision models using JAX and π€ transformers