sail-sg

Edit anything in images powered by segment-anything, ControlNet, StableDiffusion, etc. (ACM MM)

C++-based high-performance parallel environment execution engine (vectorized env) for general RL environments.

PoolFormer: MetaFormer Is Actually What You Need for Vision (CVPR 2022 Oral)

Understanding R1-Zero-Like Training: A Critical Perspective

VOLO: Vision Outlooker for Visual Recognition

Adan: Adaptive Nesterov Momentum Algorithm for Faster Optimizing Deep Models

🌾 OAT: A research-friendly framework for LLM online alignment, including reinforcement learning, preference learning, etc.

[COLM 2024] LoraHub: Efficient Cross-Task Generalization via Dynamic LoRA Composition

Masked Diffusion Transformer is the SOTA for image synthesis. (ICCV 2023)

MetaFormer Baselines for Vision (TPAMI 2024)

InceptionNeXt: When Inception Meets ConvNeXt (CVPR 2024)

NeurIPS-2021: Direct Multi-view Multi-person 3D Human Pose Estimation

CVPR'24, Official Codebase of our Paper: "Let's Think Outside the Box: Exploring Leap-of-Thought in Large Language Models with Creative Humor Generation".

A lightweight reproduction of DeepSeek-R1-Zero with indepth analysis of self-reflection behavior.

iFormer: Inception Transformer

[ICLR 2025] 🧬 RegMix: Data Mixture as Regression for Language Model Pre-training (Spotlight)

Defeating the Training-Inference Mismatch via FP16

[CVPR 2025] TreeMeshGPT: Artistic Mesh Generation with Autoregressive Tree Sequencing

[ACL 2024] The official codebase for the paper "Self-Distillation Bridges Distribution Gap in Language Model Fine-tuning".

BindDiffusion: One Diffusion Model to Bind Them All

[ICLR 2025] When Attention Sink Emerges in Language Models: An Empirical View (Spotlight)

[CVPR2023] The code for 《Position-guided Text Prompt for Vision-Language Pre-training》

[EMNLP-2024] ⚓️ Sailor: Open Language Models for South-East Asia

[NeurIPS 2024] The official implementation of paper: Chain of Preference Optimization: Improving Chain-of-Thought Reasoning in LLMs.

[ICML 2024] Agent Smith: A Single Image Can Jailbreak One Million Multimodal LLM Agents Exponentially Fast

Official implementation of Stochastic Taylor Derivative Estimator (STDE) NeurIPS2024

[NeurIPS 2023] Efficient Diffusion Policy

[NeurIPS-2024] 📈 Scaling Laws with Vocabulary: Larger Models Deserve Larger Vocabularies https://arxiv.org/abs/2407.13623

Reinforcing General Reasoning without Verifiers

The official PyTorch implementation of Fast Diffusion Model

🚢 Data Toolkit for Sailor Language Models

Automatic Functional Differentiation in JAX

[ACL 2026 (Main)] LongSpec: Long-Context Lossless Speculative Decoding with Efficient Drafting and Verification

A PyTorch implementation of Mugs proposed by our paper "Mugs: A Multi-Granular Self-Supervised Learning Framework".

[ICLR 2025] Cheating Automatic LLM Benchmarks: Null Models Achieve High Win Rates (Oral)

The official PyTorch implementation of Consistent3D (CVPR 2024)

Rethinking the Trust Region in LLM Reinforcement Learning

🔱 Sailor2: Sailing in South-East Asia with Inclusive Multilingual LLMs

Improved Few-Shot Jailbreaking Can Circumvent Aligned Language Models and Their Defenses (NeurIPS 2024)

Code for "Language Models Can Learn from Verbal Feedback Without Scalar Rewards"

The official repository for the paper "From Zero to Hero: Examining the Power of Symbolic Tasks in Instruction Tuning".

Code for "Variational Reasoning for Language Models"

AnyDoor: Test-Time Backdoor Attacks on Multimodal Large Language Models

A JAX library for Density Functional Theory.

The official implementation of paper: SimLayerKV: A Simple Framework for Layer-Level KV Cache Reduction.

Optimizing Anytime Reasoning via Budget Relative Policy Optimization

Exchange correlation functionals translated from libxc to jax

A JAX-based Differentiable Density Functional Theory Framework for Materials

[CVPR 2026] TeamHOI: Learning a Unified Policy for Cooperative Human-Object Interactions with Any Team Size

[ICLR 2025] A Closer Look at Machine Unlearning for Large Language Models

The official repository of paper "ScaleLong: Towards More Stable Training of Diffusion Model via Scaling Network Long Skip Connection" (NeurIPS 2023)