Zhaokun Zhou

Hetero-RL. Group Expectation Policy Optimization for Heterogeneous Reinforcement Learning

DeepSeek-Math. DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models

WISE. [ICML 2026🔥] WISE: A World Knowledge-Informed Semantic Evaluation for Text-to-Image Generation

QSD-Transformer. Offical implementation of "Quantized Spike-driven Transformer" (ICLR2025)

FlashMLA. FlashMLA: Efficient Multi-head Latent Attention Kernels

MoH. MoH: Multi-Head Attention as Mixture-of-Head Attention

SpikeYOLO. Offical implementation of "Integer-Valued Training and Spike-Driven Inference Spiking Neural Network for High-performance and Energy-efficient Object Detection" (ECCV2024 Best Paper Candidate)

MetaLA. Offical implementation of "MetaLA: Unified Optimal Linear Approximation to Softmax Attention Map" (NeurIPS2024 Oral)

Spike-Driven-Transformer-V3. Offical implementation of "Scaling Spike-driven Transformer with Efficient Spike Firing Approximation Training" (IEEE T-PAMI2025)

IC-Light. More relighting!

transformer_vq. Official implementation of 'Transformer-VQ: Linear-Time Transformers via Vector Quantization' (ICLR 2024)

SpikedAttention. This is simple code of SpikedAttention (Neurips 2024)

LLaVA-CoT. [ICCV 2025] LLaVA-CoT, a visual language model capable of spontaneous, systematic reasoning

ai-for-grant-writing. A curated list of resources for using LLMs to develop more competitive grant applications.

ViewCrafter. [TPAMI 2025] ViewCrafter: Taming Video Diffusion Models for High-fidelity Novel View Synthesis

AMBER. An LLM-free Multi-dimensional Benchmark for Multi-modal Hallucination Evaluation

outlier-free-transformers. Python

AesExpert. [ACMMM 2024] AesExpert: Towards Multi-modality Foundation Model for Image Aesthetics Perception

AesBench. An expert benchmark aiming to comprehensively evaluate the aesthetic perception capacities of MLLMs.

Uniaa. Unified Multi-modal IAA Baseline and Benchmark

mamba. Mamba SSM architecture

QKFormer. Offical code of "QKFormer: Hierarchical Spiking Transformer using Q-K Attention"

MotionInversion. [SIGGRAPH 2025] Official implementation of 'Motion Inversion For Video Customization'

Spike-Driven-Transformer-V2. Offical implementation of "Spike-driven Transformer V2: Meta Spiking Neural Network Architecture Inspiring the Design of Next-generation Neuromorphic Chips" (ICLR2024)

Masked-Spiking-Transformer. [ICCV-23 + IJCV-26] Masked Spiking Transformer

SGLFormer. Spiking Global-Local Fusion Transformer

SGLFormer. Spiking Global-Local Fusion Transformer

ProLLaMA. A Protein Large Language Model for Multi-Task Protein Language Processing

ECDFormer. 【Nature Computational Science 2025🔥】Deep peak property learning for efficient chiral molecules ECD spectra prediction

mamba-minimal. Simple, minimal implementation of the Mamba SSM in one file of PyTorch.

MiniGPT-4. Open-sourced codes for MiniGPT-4 and MiniGPT-v2 (https://minigpt-4.github.io, https://minigpt-v2.github.io/)

Qwen-VL. The official repo of Qwen-VL (通义千问-VL) chat & pretrained large vision language model proposed by Alibaba Cloud.

PiCO. [ICLR'25] PiCO: Peer Review in LLMs based on the Consistency Optimization, https://arxiv.org/pdf/2402.01830

Awesome-Spiking-Neural-Networks. A paper list of spiking neural networks, including papers, codes, and related websites. 本仓库收集脉冲神经网络相关的顶会顶刊以及CNS论文和代码，正在持续更新中。

HARDVS. [AAAI-2024, IJCV-2026] HARDVS: Revisiting Human Activity Recognition with Dynamic Vision Sensors

MoE-LLaVA. 【TMM 2025🔥】 Mixture-of-Experts for Large Vision-Language Models

Aestheval. Code for the paper "Understanding Aesthetics with Language: A Photo Critique Dataset for Aesthetic Assessment"

BAID. Python

Vim. [ICML 2024] Vision Mamba: Efficient Visual Representation Learning with Bidirectional State Space Model

VMamba. VMamba: Visual State Space Models，code is based on mamba

SSAH-adversarial-attack. Code for the paper "Frequency-driven Imperceptible Adversarial Attack on Semantic Similarity"

LM4VisualEncoding. [ICLR 2024 (Spotlight)] "Frozen Transformers in Language Models are Effective Visual Encoder Layers"

xtuner. A Next-Generation Training Engine Built for Ultra-Large MoE Models

Vary. [ECCV 2024] Official code implementation of Vary: Scaling Up the Vision Vocabulary of Large Vision Language Models.

Q-Bench. An archived version of Q-Bench. We will make updates in https://github.com/q-future/Q-Bench in the future.

Image-Color-Aesthetics-and-Quality-Assessment. 🔥[ICCV 2023, Official Code] for paper "Thinking Image Color Aesthetics Assessment: Models, Datasets and Benchmarks". Official Weights and Demos provided. 首个面向图像色彩主观美学评估的数据集、算法和benchmark.

StableLLAVA. Official repo for StableLLAVA

Otter. 🦦 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.

Video-LLaVA. 【EMNLP 2024🔥】Video-LLaVA: Learning United Visual Representation by Alignment Before Projection

Image-Aesthetics-and-Quality-Assessment. 🔥[ACMMM 2023, Official Code] for paper "EAT: An Enhancer for Aesthetics-Oriented Transformers". Official Weights and Demos provided. 目前是地表最强开源美学评估模型之一.

EMCL. [NeurIPS 2022 Spotlight] Expectation-Maximization Contrastive Learning for Compact Video-and-Language Representations

GraphMotion. [NeurIPS 2023] Act As You Wish: Fine-Grained Control of Motion Diffusion Model with Hierarchical Semantic Graphs

MQ-Det. Official PyTorch implementation of "Multi-modal Queried Object Detection in the Wild" (accepted by NeurIPS 2023)

Dynamic-Token-Pruning. Official Pytorch implementation of Dynamic-Token-Pruning (ICCV2023)

DynamicViT. [NeurIPS 2021] [T-PAMI] DynamicViT: Efficient Vision Transformers with Dynamic Token Sparsification

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

Medusa. Medusa: Simple Framework for Accelerating LLM Generation with Multiple Decoding Heads

Awesome-Multimodal-Large-Language-Models. :sparkles::sparkles:Latest Advances on Multimodal Large Language Models

baby-llama2-chinese. 用于从头预训练+SFT一个小参数量的中文LLaMa2的仓库；24G单卡即可运行得到一个具备简单中文问答能力的chat-llama2.

loss-landscape. Code for visualizing the loss landscape of neural nets

EmbodiedGPT_Pytorch. Python

MADAv2. MADAv2: Advanced Multi-Anchor Based Active Domain Adaptation Segmentation

Spike-Driven-Transformer. Offical implementation of "Spike-driven Transformer" (NeurIPS2023)

Fast-SNN. Python

RRHF. [NIPS2023] RRHF & Wombat

ChatLaw. ChatLaw：A Powerful LLM Tailored for Chinese Legal. 中文法律大模型

Awesome-LLM. Awesome-LLM: a curated list of Large Language Model

NeuroCLIP. Python

MAE-Lite. Official implement for ICML2023 paper: "A Closer Look at Self-Supervised Lightweight Vision Transformers"

ConvNeXt-V2. Code release for ConvNeXt V2 model

VanillaNet. Python

SAVC. [CVPR 2023] Learning with Fantasy: Semantic-Aware Virtual Contrastive Constraint for Few-Shot Class-Incremental Learning

qlora. QLoRA: Efficient Finetuning of Quantized LLMs

PointGPT. [NeurIPS 2023] PointGPT: Auto-regressively Generative Pre-training from Point Clouds

Spikingformer. Spikingformer: A Key Foundation Model for Spiking Neural Networks (AAAI 2026)

Spikingformer-CML. Enhancing the Performance of Transformer-based Spiking Neural Networks by SNN-optimized Downsampling with Precise Gradient Backpropagation

Swin-MAE. Pytorch implementation of Swin MAE https://arxiv.org/abs/2212.13805

Compact-Transformers. Escaping the Big Data Paradigm with Compact Transformers, 2021 (Train your Vision Transformers in 30 mins on CIFAR-10 with a single GPU!)

LLM-in-Vision. Recent LLM-based CV and related works. Welcome to comment/contribute!

FCViT. A Close Look at Spatial Modeling: From Attention to Convolution

hivit. Jupyter Notebook

Parallel-Spiking-Neuron. Python

SNNCKA. Pytorch implementation of ANN-SNN representation similarity analysis (TMLR 2023)

ml-cvnets. CVNets: A library for training computer vision networks

Spike-Element-Wise-ResNet. Deep Residual Learning in Spiking Neural Networks

efficientvit. Efficient vision foundation models for high-resolution generation and perception.

segment-anything. The repository provides code for running inference with the SegmentAnything Model (SAM), links for downloading the trained model checkpoints, and example notebooks that show how to use the model.

RWKV-LM. RWKV (pronounced RwaKuv) is an RNN with great LLM performance, which can also be directly trained like a GPT transformer (parallelizable). We are at RWKV-7 "Goose". So it's combining the best of RNN and transformer - great performance, linear time, constant space (no kv-cache), fast training, infinite ctx_len, and free sentence embedding.

Efficient-Computing. Efficient computing methods developed by Huawei Noah's Ark Lab

ConvMAE. ConvMAE: Masked Convolution Meets Masked Autoencoders

SparK. [ICLR'23 Spotlight🔥] The first successful BERT/MAE-style pretraining on any convolutional network; Pytorch impl. of "Designing BERT for Convolutional Networks: Sparse and Hierarchical Masked Modeling"

spikformer. ICLR 2023, Spikformer: When Spiking Neural Network Meets Transformer

CAE. This is a PyTorch implementation of “Context AutoEncoder for Self-Supervised Representation Learning"

detectron2. Detectron2 is a platform for object detection, segmentation and other visual recognition tasks.

MIMDet. [ICCV 2023] You Only Look at One Partial Sequence

Attention-driven-Masking-and-Throwing. Python

SNN-Neural-Similarity-Static. Deep Spiking Neural Networks with High Representation Similarity Model Visual Pathways of Macaque and Mouse

mae. PyTorch implementation of MAE https//arxiv.org/abs/2111.06377

syops-counter. SyOPs counter for spiking neural networks

Q-ViT. The official implementation of the NeurIPS 2022 paper Q-ViT.

rebiber. A simple tool to update bib entries with their official information (e.g., DBLP or the ACL anthology).

Awesome-Transformer-Attention. An ultimately comprehensive paper list of Vision Transformer/Attention, including papers, codes, and related websites

MAE-pytorch. Unofficial PyTorch implementation of Masked Autoencoders Are Scalable Vision Learners

Point-MAE. [ECCV2022] Masked Autoencoders for Point Cloud Self-supervised Learning

spikingjelly. SpikingJelly is an open-source deep learning framework for Spiking Neural Network (SNN) based on PyTorch.

Swin-Transformer. This is an official implementation for "Swin Transformer: Hierarchical Vision Transformer using Shifted Windows".

focal_loss_pytorch. A PyTorch Implementation of Focal Loss.

auto-feature-extraction-method-for-e-tongue. Python

slam_gmapping. http://www.ros.org/wiki/slam_gmapping