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

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

Official Code for "SVD-LLM: Truncation-aware Singular Value Decomposition for Large Language Model Compression"

For releasing code related to compression methods for transformers, accompanying our publications

PyTorch native quantization and sparsity for training and inference

QServe: W4A8KV4 Quantization and System Co-design for Efficient LLM Serving

Official repository for "AM-RADIO: Reduce All Domains Into One"

The official implementation of the DAC 2024 paper GQA-LUT

Efficiently Fine-Tune 100+ LLMs in WebUI (ACL 2024)

Training LLMs with QLoRA + FSDP

The official Meta Llama 3 GitHub site

[ICML2024 (Oral)] Official PyTorch implementation of DoRA: Weight-Decomposed Low-Rank Adaptation

Code for QuaRot, an end-to-end 4-bit inference of large language models.

ReFT: Representation Finetuning for Language Models

A simple and effective LLM pruning approach.

PyTorch code for Vision Transformers training with the Self-Supervised learning method DINO

This is a project page for continuous 3D words

awesome llm plaza: daily tracking all sorts of awesome topics of llm, e.g. llm for coding, robotics, reasoning, multimod etc.

Official implementation of "DoRA: Weight-Decomposed Low-Rank Adaptation"

Modeling, training, eval, and inference code for OLMo

The code and data for the GPT-4 based benchmark in the vicuna blog post

Instruct-tune LLaMA on consumer hardware

[NeurIPS'23 Oral] Visual Instruction Tuning (LLaVA) built towards GPT-4V level capabilities and beyond.

Official implementation of "Parameter-Efficient Orthogonal Finetuning via Butterfly Factorization"

TensorRT-LLM provides users with an easy-to-use Python API to define Large Language Models (LLMs) and build TensorRT engines that contain state-of-the-art optimizations to perform inference efficie…

🤗 PEFT: State-of-the-art Parameter-Efficient Fine-Tuning.

Code for loralib, an implementation of "LoRA: Low-Rank Adaptation of Large Language Models"

The official implementation of the EMNLP 2023 paper LLM-FP4

A curated list of awesome vision and language resources (still under construction... stay tuned!)