dlwh

vllm

A high-throughput and memory-efficient inference and serving engine for LLMs

mlx

MLX: An array framework for Apple silicon

marker

Convert PDF to markdown quickly with high accuracy

FlexGen

Running large language models on a single GPU for throughput-oriented scenarios.

PokemonRedExperiments

Playing Pokemon Red with Reinforcement Learning

gpt-fast

Simple and efficient pytorch-native transformer text generation in <1000 LOC of python.

brax

Massively parallel rigidbody physics simulation on accelerator hardware.

big_vision

Official codebase used to develop Vision Transformer, SigLIP, MLP-Mixer, LiT and more.

huggingface_hub

The official Python client for the Huggingface Hub.

axlearn

An Extensible Deep Learning Library

optax

Optax is a gradient processing and optimization library for JAX.

alpaca_eval

An automatic evaluator for instruction-following language models. Human-validated, high-quality, cheap, and fast.

deduplicate-text-datasets

lilac

Curate better data for LLMs

jaxopt

Hardware accelerated, batchable and differentiable optimizers in JAX.

safari

Convolutions for Sequence Modeling

BricksLLM

🔒 Enterprise-grade API gateway that helps you monitor and impose cost or rate limits per API key. Get fine-grained access control and monitoring per user, application, or environment. Supports OpenAI, Azure OpenAI, Anthropic, vLLM, and open-source LLMs.

enroot

A simple yet powerful tool to turn traditional container/OS images into unprivileged sandboxes.

Megatron-LLM

distributed trainer for LLMs

torchdim

Named tensors with first-class dimensions for PyTorch

pyxis

Container plugin for Slurm Workload Manager

gears

A strawman for a low-level async library in Scala 3.

grain

universal_pathlib

pathlib api extended to use fsspec backends

voltron-robotics

Voltron: Language-Driven Representation Learning for Robotics

EasyDeL

Accelerate your training with this open-source library. Optimize performance with streamlined training and serving options with JAX. 🚀

lm-human-preference-details

RLHF implementation details of OAI's 2019 codebase

lovely-jax

JAX Arrays for human consumption

xpk

xpk (Accelerated Processing Kit, pronounced x-p-k,) is a software tool to help Cloud developers to orchestrate training jobs on accelerators such as TPUs and GPUs on GKE.

levanter-midi

Experiments with Jax and TPUs for Foundation Models - modified tokens for lakh