Making large AI models cheaper, faster and more accessible

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

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

🦦 Otter, a multi-modal model based on OpenFlamingo (open-sourced version of DeepMind's Flamingo), trained on MIMIC-IT and showcasing improved instruction-following and in-context learning ability.

Code and models for NExT-GPT: Any-to-Any Multimodal Large Language Model

[CVPR2024 Highlight][VideoChatGPT] ChatGPT with video understanding! And many more supported LMs such as miniGPT4, StableLM, and MOSS.

SuperCLUE: 中文通用大模型综合性基准 | A Benchmark for Foundation Models in Chinese

Chronos: Pretrained (Language) Models for Probabilistic Time Series Forecasting

EVA Series: Visual Representation Fantasies from BAAI

DeepSeek-VL: Towards Real-World Vision-Language Understanding

Images to inference with no labeling (use foundation models to train supervised models).

Prompt Learning for Vision-Language Models (IJCV'22, CVPR'22)

Emu Series: Generative Multimodal Models from BAAI

Awesome things about LLM-powered agents. Papers / Repos / Blogs / ...

An automatic evaluator for instruction-following language models. Human-validated, high-quality, cheap, and fast.

[ECCV2024] Video Foundation Models & Data for Multimodal Understanding

Lag-Llama: Towards Foundation Models for Probabilistic Time Series Forecasting

Janus-Series: Unified Multimodal Understanding and Generation Models

日本語LLMまとめ - Overview of Japanese LLMs

A general representation model across vision, audio, language modalities. Paper: ONE-PEACE: Exploring One General Representation Model Toward Unlimited Modalities