OCR Annotations from Amazon Textract for Industry Documents Library

OLMoE: Open Mixture-of-Experts Language Models

A curated list of awesome TikZ packages and resources

Oracle Bone Script data collected by VLRLab of HUST

🦉 OWL: Optimized Workforce Learning for General Multi-Agent Assistance in Real-World Task Automation

[NeurIPS'24]Efficient and accurate memory saving method towards W4A4 large multi-modal models.

Extend OpenRLHF to support LMM RL training for reproduction of DeepSeek-R1 on multimodal tasks.

MoBA: Mixture of Block Attention for Long-Context LLMs

USP: Unified (a.k.a. Hybrid, 2D) Sequence Parallel Attention for Long Context Transformers Model Training and Inference

Gradio WebUI for creators and developers, featuring key TTS (Edge-TTS, kokoro) and zero-shot Voice Cloning (E2 & F5-TTS, CosyVoice), with Whisper audio processing, YouTube download, Demucs vocal is…

Automating the Search for Artificial Life with Foundation Models!

提供历代书法绘画作品的描述信息和图片数据文件下载，为书画AI研究提供训练数据，也可以用于其他和传统艺术相关的科研工作。

An Open Large Reasoning Model for Real-World Solutions

Moshi is a speech-text foundation model and full-duplex spoken dialogue framework. It uses Mimi, a state-of-the-art streaming neural audio codec.

AllenAI's post-training codebase

Low Precision Arithmetic Simulation in PyTorch

Pruning the VLLMs

LLaVA-CoT, a visual language model capable of spontaneous, systematic reasoning

Making large AI models cheaper, faster and more accessible

Evaluating text-to-image/video/3D models with VQAScore

Fast Multimodal LLM on Mobile Devices

High-speed Large Language Model Serving for Local Deployment

[NeurIPS 2024 Oral🔥] DuQuant: Distributing Outliers via Dual Transformation Makes Stronger Quantized LLMs.

Next-Token Prediction is All You Need

Data processing for and with foundation models! 🍎 🍋 🌽 ➡️ ➡️🍸 🍹 🍷

Flux diffusion model implementation using quantized fp8 matmul & remaining layers use faster half precision accumulate, which is ~2x faster on consumer devices.

EfficientQAT: Efficient Quantization-Aware Training for Large Language Models