AutoAWQ implements the AWQ algorithm for 4-bit quantization with a 2x speedup during inference. Documentation:

The most powerful and modular diffusion model GUI, api and backend with a graph/nodes interface.

Code repo for the paper "SpinQuant LLM quantization with learned rotations"

Puzzles for learning Triton

Production ready LLM model compression/quantization toolkit with accelerated inference support for both cpu/gpu via HF, vLLM, and SGLang.

Open source replication of Anthropic's Crosscoders for Model Diffing

Create feature-centric and prompt-centric visualizations for sparse autoencoders (like those from Anthropic's published research).

🤗 Diffusers: State-of-the-art diffusion models for image and audio generation in PyTorch and FLAX.

A framework for the evaluation of autoregressive code generation language models.

FP16xINT4 LLM inference kernel that can achieve near-ideal ~4x speedups up to medium batchsizes of 16-32 tokens. Version with scale finetuning.

Consistency Distilled Diff VAE

A Compressed Stable Diffusion for Efficient Text-to-Image Generation [ICCV'23 Demo] [ICML'23 Workshop]

LLM sparse training code for MosaicML foundation models

Code for fine-tuning Platypus fam LLMs using LoRA

Code for the paper "SparseGPT: Massive Language Models Can Be Accurately Pruned in One-Shot".

Libraries for applying sparsification recipes to neural networks with a few lines of code, enabling faster and smaller models

Detectron2 is a platform for object detection, segmentation and other visual recognition tasks.

Training GNNs on heterophilous graphs.