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

MARS5 speech model (TTS) from CAMB.AI

A distributed approximate nearest neighborhood search (ANN) library which provides a high quality vector index build, search and distributed online serving toolkits for large scale vector search sc…

[ECCV 2024] MOFA-Video: Controllable Image Animation via Generative Motion Field Adaptions in Frozen Image-to-Video Diffusion Model.

A framework for few-shot evaluation of language models.

LangEvals aggregates various language model evaluators into a single platform, providing a standard interface for a multitude of scores and LLM guardrails, for you to protect and benchmark your LLM…

Lite & Super-fast re-ranking for your search & retrieval pipelines. Supports SoTA Listwise and Pairwise reranking based on LLMs and cross-encoders and more. Created by Prithivi Da, open for PRs & C…

RAGElo is a set of tools that helps you selecting the best RAG-based LLM agents by using an Elo ranker

Examples in the MLX framework

BigCodeBench: The Next Generation of HumanEval

A diffusers pipeline for zero shot stylised portrait creation

Go ahead and axolotl questions

End-to-End Speech Processing Toolkit

An easy-to-use LLMs quantization package with user-friendly apis, based on GPTQ algorithm.

[CVPR 2024] Official RT-DETR (RTDETR paddle pytorch), Real-Time DEtection TRansformer, DETRs Beat YOLOs on Real-time Object Detection. 🔥 🔥 🔥

WebUI for Fine-Tuning and Self-hosting of Open-Source Large Language Models for Coding

Minimal, clean code for the Byte Pair Encoding (BPE) algorithm commonly used in LLM tokenization.

Port of Andrej Karpathy's minbpe to Rust

Recipes to train reward model for RLHF.

Port of OpenAI's Whisper model in C/C++

[arXiv 2024] Follow-Your-Click: This repo is the official implementation of "Follow-Your-Click: Open-domain Regional Image Animation via Short Prompts"

💥 Fast State-of-the-Art Tokenizers optimized for Research and Production

YaFSDP: Yet another Fully Sharded Data Parallel

[CCS'24] A dataset consists of 15,140 ChatGPT prompts from Reddit, Discord, websites, and open-source datasets (including 1,405 jailbreak prompts).

A programming framework for agentic AI. Discord: https://aka.ms/autogen-dc. Roadmap: https://aka.ms/autogen-roadmap

active angel investors in developer tools!

Qwen2 is the large language model series developed by Qwen team, Alibaba Cloud.