HarmBench: A Standardized Evaluation Framework for Automated Red Teaming and Robust Refusal

Deep Research Agent CognitiveKernel-Pro from Tencent AI Lab. Paper: https://arxiv.org/pdf/2508.00414

slime is an LLM post-training framework for RL Scaling.

verl: Volcano Engine Reinforcement Learning for LLMs

GLM-4.5: Agentic, Reasoning, and Coding (ARC) Foundation Models

GLM-4.5V and GLM-4.1V-Thinking: Towards Versatile Multimodal Reasoning with Scalable Reinforcement Learning

An Easy-to-use, Scalable and High-performance RLHF Framework based on Ray (PPO & GRPO & REINFORCE++ & vLLM & Ray & Dynamic Sampling & Async Agentic RL)

Official Repository of ACL 2025 paper OmniAlign-V: Towards Enhanced Alignment of MLLMs with Human Preference

EVA Series: Visual Representation Fantasies from BAAI

This is the code repository of our submission: Understanding the Dark Side of LLMs’ Intrinsic Self-Correction.

[NeurIPS 2024 Best Paper Award][GPT beats diffusion🔥] [scaling laws in visual generation📈] Official impl. of "Visual Autoregressive Modeling: Scalable Image Generation via Next-Scale Prediction". A…

MoVQGAN - model for the image encoding and reconstruction

[CVPR 2024] Not All Prompts Are Secure: A Switchable Backdoor Attack Against Pre-trained Vision Transfomers

Official repository for paper "MagicMan: Generative Novel View Synthesis of Humans with 3D-Aware Diffusion and Iterative Refinement"

Official Repository of ChartX & ChartVLM: A Versatile Benchmark and Foundation Model for Complicated Chart Reasoning

[IEEE TPAMI 2025] This repository is the official implementation of the paper "VisionUnite: A Vision-Language Foundation Model for Ophthalmology Enhanced with Clinical Knowledge"

Repository for Meta Chameleon, a mixed-modal early-fusion foundation model from FAIR.

PyTorch implementation of MAR+DiffLoss https://arxiv.org/abs/2406.11838

Paper list about multimodal and large language models, only used to record papers I read in the daily arxiv for personal needs.

Ongoing research training transformer models at scale

Source code for the ICDM paper ``HINT: Healthy Influential-Noise based Training to Defend against Data Poisoning Attacks"

This is a repo of "3D adversarial sample".

PyTorch implementation of Expectation over Transformation

Code for Friendly Noise against Adversarial Noise: A Powerful Defense against Data Poisoning Attacks (NeurIPS 2022)

A unified framework for 3D content generation.

The most critical downside to AI is that its inefficiency is directly related to its data quality. Presently there are very few methods available that are used to protect the data from being attack…