A family of compressed models obtained via pruning and knowledge distillation

Open weights language model from Google DeepMind, based on Griffin.

GUICourse: From General Vision Langauge Models to Versatile GUI Agents

Sparse Backpropagation for Mixture-of-Expert Training

The code for "TokenPacker: Efficient Visual Projector for Multimodal LLM".

Together Mixture-Of-Agents (MoA) – 65.1% on AlpacaEval with OSS models

The official evaluation suite and dynamic data release for MixEval.

[CVPR2024 Highlight] VBench - We Evaluate Video Generation

Skywork-MoE: A Deep Dive into Training Techniques for Mixture-of-Experts Language Models

📺 An End-to-End Solution for High-Resolution and Long Video Generation Based on Transformer Diffusion

The homepage of OneBit model quantization framework.

DeepSeek-V2: A Strong, Economical, and Efficient Mixture-of-Experts Language Model

Safe RLHF: Constrained Value Alignment via Safe Reinforcement Learning from Human Feedback

Evaluation framework for paper "VisualWebBench: How Far Have Multimodal LLMs Evolved in Web Page Understanding and Grounding?"

An LLM-based Web Navigating Agent (KDD'24)

Reaching LLaMA2 Performance with 0.1M Dollars

Web-grounded natural language instructions

[ICML2024]Adaptive decoding balances the diversity and coherence of open-ended text generation.

VisualWebArena is a benchmark for multimodal agents.

Code repo for "WebArena: A Realistic Web Environment for Building Autonomous Agents"

The model, data and code for the visual GUI Agent SeeClick

Grok open release

Open-Sora: Democratizing Efficient Video Production for All

VILA - a multi-image visual language model with training, inference and evaluation recipe, deployable from cloud to edge (Jetson Orin and laptops)

[ICML'24] SeeAct is a system for generalist web agents that autonomously carry out tasks on any given website, with a focus on large multimodal models (LMMs) such as GPT-4V(ision).