A generative world for general-purpose robotics & embodied AI learning.

RWKV is an RNN with transformer-level LLM performance. It can be directly trained like a GPT (parallelizable). So it's combining the best of RNN and transformer - great performance, fast inference,…

InternLM-XComposer2.5-OmniLive: A Comprehensive Multimodal System for Long-term Streaming Video and Audio Interactions

An open-source implementation for training LLaVA-NeXT.

Simple implementation of Speculative Sampling in NumPy for GPT-2.

This is the official repo for Debiasing Large Visual Language Models, including a Post-Hoc debias method and Visual Debias Decoding strategy.

[ICML 2024] Language Models Represent Beliefs of Self and Others

This repository contains the code for the paper: Beyond Chain-of-Thought, Effective Graph-of-Thought Reasoning in Language Models

Code for the "Long Context Needs Some R&R" paper.