Large-scale Self-supervised Pre-training Across Tasks, Languages, and Modalities

Evals is a framework for evaluating LLMs and LLM systems, and an open-source registry of benchmarks.

Finetune Llama 3, Mistral, Phi & Gemma LLMs 2-5x faster with 80% less memory

Structured Text Generation

Accessible large language models via k-bit quantization for PyTorch.

Perceptual video quality assessment based on multi-method fusion.

Unified framework for building enterprise RAG pipelines with small, specialized models

Easily use and train state of the art late-interaction retrieval methods (ColBERT) in any RAG pipeline. Designed for modularity and ease-of-use, backed by research.

RAG AutoML Tool - Find optimal RAG pipeline for your own data.

Efficient Retrieval Augmentation and Generation Framework

OSWorld: Benchmarking Multimodal Agents for Open-Ended Tasks in Real Computer Environments

Code for 'LLM2Vec: Large Language Models Are Secretly Powerful Text Encoders'

A Python script to control Linak standing desks.

ExtractThinker is a Document Intelligence library for LLMs, offering ORM-style interaction for flexible and powerful document workflows.

a little GUI tool to render visualization videos of audio files

This repository contains the experimental PyTorch native float8 training UX

Tools for evaluation of RAG Chat Apps using Azure AI Evaluate SDK and OpenAI

The official implementation of the EMNLP 2023 paper LLM-FP4

Client-side toolkit for using large language models, including where self-hosted

PyTorch extension for emulating FP8 data formats on standard FP32 Xeon/GPU hardware.

Faster Pytorch bitsandbytes 4bit fp4 nn.Linear ops

Adaptive floating-point based numerical format for resilient deep learning

Generic floating-point types in Python