Yanbin Liu

trl. Train transformer language models with reinforcement learning.

skills. Give your agents the power of the Hugging Face ecosystem

awesome-courses. I will collect a list of (open) courses first for AI, CS and Math. High-quality knowledge matters! Try to update everyday!

SmartBench. SmartBench

Deep-Learning-Based-Anomaly-Detection.

SSG. [CVPR 2026 Oral] Guiding a Diffusion Model by Swapping Its Tokens

camouflaged-vlm. Open-Vocabulary Camouflaged Object Segmentation with Cascaded Vision Language Models

AI-Research-SKILLs. Comprehensive open-source library of AI research and engineering skills for any AI model. Package the skills and your claude code/codex/gemini agent will be an AI research agent with full horsepower. Maintained by Orchestra Research.

OTT-Vid. Official code for "OTT-Vid: Optimal Transport Temporal Token Compression for Video Large Language Models"

Awesome-Attention-Sink. 🚀 First survey on Attention Sink in Transformers — 200+ papers on utilization, interpretation, and mitigation.

FasterVLM. Official code for paper: [CLS] Attention is All You Need for Training-Free Visual Token Pruning: Make VLM Inference Faster.

LLaVA-PruMerge. LLaVA-PruMerge: Adaptive Token Reduction for Efficient Large Multimodal Models

VLMEvalKit. Open-source evaluation toolkit of large multi-modality models (LMMs), support 220+ LMMs, 80+ benchmarks

DVS. ICML 2026: Information-geometric adaptive sampling for graph diffusion

drifting. Python

IMBA-Loss. [ICCV 2025] Official Implementation of the Paper "Imbalance in Balance: Online Concept Balancing in Generation Models".

mllms_know. [ICLR'25] Official code for the paper 'MLLMs Know Where to Look: Training-free Perception of Small Visual Details with Multimodal LLMs'

LLMxFM_ICRL. Code for the "Can LLMs Reason Over Non-Text Modalities in a Training-Free Manner? A Case Study with In-Context Representation Learning" paper

evo-memory. Code to train and evaluate Neural Attention Memory Models to obtain universally-applicable memory systems for transformers.

MetaSPO. Python

Binoculars. [ICML 2024] Binoculars: Zero-Shot Detection of LLM-Generated Text

fast-detect-gpt. Code base for ICLR 2024 "Fast-DetectGPT: Efficient Zero-Shot Detection of Machine-Generated Text via Conditional Probability Curvature".

tuning-contribution. Official codebase for TuCo: Measuring the Contribution of Fine-Tuning to Individual Responses of LLMs (ICML 2025).

prompts.chat. f.k.a. Awesome ChatGPT Prompts. Share, discover, and collect prompts from the community. Free and open source — self-host for your organization with complete privacy.

chatgpt-prompts-for-academic-writing. This list of writing prompts covers a range of topics and tasks, including brainstorming research ideas, improving language and style, conducting literature reviews, and developing research plans.

ClavaDDPM. [NeurIPS 2024] Official implementation of "ClavaDDPM:Multi-relational Data Synthesis with Cluster-guided Diffusion Models"

GruM. Official Code Repository for the paper "Graph Generation with Diffusion Mixture" (ICML 2024).

denoising-diffusion-pytorch. Implementation of Denoising Diffusion Probabilistic Model in Pytorch

ICML2025-TCPA. Python

Academic-LaTeX-Writing-Submission-Checklist-. This checklist is designed to help you systematically prepare and polish academic papers for top conferences and journals (e.g., ICML, NeurIPS, CVPR). It incorporates widely recommended best practices, formatting standards, and common reviewer expectations.

segment-anything-annotator. We developed a python UI based on labelme and segment-anything for pixel-level annotation. It support multiple masks generation by SAM(box/point prompt), efficient polygon modification and category record. We will add more features (such as incorporating CLIP-based methods for category proposal and VOS methods for video datasets

Segment-Anything-CLIP. Connecting segment-anything's output masks with the CLIP model; Awesome-Segment-Anything-Works

SAMRS. The official repo for [NeurIPS'23] "SAMRS: Scaling-up Remote Sensing Segmentation Dataset with Segment Anything Model"

RSRefSeg. This is the pytorch implement of the paper "RSRefSeg: Referring Remote Sensing Image Segmentation with Foundation Models"

RobustSAM. RobustSAM: Segment Anything Robustly on Degraded Images (CVPR 2024 Highlight)

SkySense-O. [CVPR 2025] This is a model aggregated with CLIP and SAM version of SkySense for remote sensing interpretation described in SkySense-O: Towards Open-World Remote Sensing Interpretation with Vision-Centric Visual-Language Modeling.

GroundingDINO. [ECCV 2024] Official implementation of the paper "Grounding DINO: Marrying DINO with Grounded Pre-Training for Open-Set Object Detection"

GeoPix. [GRSM] Project Page for "GeoPix: Multi-Modal Large Language Model for Pixel-level Image Understanding in Remote Sensing"

GeoPixel. GeoPixel: A Pixel Grounding Large Multimodal Model for Remote Sensing is specifically developed for high-resolution remote sensing image analysis, offering advanced multi-target pixel grounding capabilities.

SegEarth-OV. [CVPR 2025 Oral] SegEarth-OV: Towards Training-Free Open-Vocabulary Segmentation for Remote Sensing Images

open-sentinel-map. Dataset download instructions.

BayesianLM. [NeurIPS 2024] "Bayesian-Guided Label Mapping for Visual Reprogramming". Official Website: https://github.com/tmlr-group/BayesianLM

BayesianLM. [NeurIPS 2024 Oral] "Bayesian-Guided Label Mapping for Visual Reprogramming"

SAVs. Official Codebase for "Generative Multimodal Model Features Are Discriminative Vision-Language Classifiers"

OpenCity. "OpenCity: Open Spatio-Temporal Foundation Models for Traffic Prediction"

synthehicle. [WACVW 2023] A massive synthetic dataset for 3D multi-target multi-camera tracking and segmentation.

torch-LLM4SGG. Official PyTorch implementation Source code for LLM4SGG: Large Language Models for Weakly Supervised Scene Graph Generation, accepted at CVPR 2024

multimodal-icl. Code repo for paper Can Multimodal Large Language Models Truly Perform Multimodal In-Context Learning? WACV 2025

DiSA. Official Implementation of Diffusion Step Annealing (DiSA) in Autoregressive Image Generation

fewshotgraph. Awesome Few-Shot Learning on Graphs

imu2clip. Code repository for IMU2CLIP(https//arxiv.org/pdf/2210.14395.pdf)

COMODO. [UbiComp/IMWUT '26] Official Repo for COMODO: Cross-Modal Video-to-IMU Distillation for Efficient Egocentric Human Activity Recognition

SelfPAB. Large-Scale Pre-Training for Dual-Accelerometer Human Activity Recognition

Self-Supervised-Learning-HAR. Python

causal. notebooks on methods for causal inference

EuroSAT. EuroSAT: Land Use and Land Cover Classification with Sentinel-2

gallop. Adaptation of vision-language models (CLIP) to downstream tasks using local and global prompts.

SymDPO. We have developed Symbol Demonstration Direct Preference Optimization (SymDPO) and validating its effectiveness across multiple benchmarks.

CASS. [CVPR 2025] Official Pytorch Code for Distilling Spectral Graph for Object-Context Aware Open-Vocabulary Semantic Segmentation

reevo. [NeurIPS 2024] ReEvo: Large Language Models as Hyper-Heuristics with Reflective Evolution

EoH. Evolution of Heuristics

CMuST. [NeurIPS 2024 Oral] Repository of the CMuST paper: "Get Rid of Isolation: A Continuous Multi-task Spatio-Temporal Learning Framework"

AustraliaFires. Exploratory analysis of wild-fires in Australia & a machine learning approach for wildfire modeling in Google Earth Engine

WeCLIP. CVPR2024

Time-Series-Library. A Library for Advanced Deep Time Series Models for General Time Series Analysis.

vmf-lib. Python

SeafloorAI. Python

FastChat. An open platform for training, serving, and evaluating large language models. Release repo for Vicuna and Chatbot Arena.

llama-models. Utilities intended for use with Llama models.

CVinW_Readings. A collection of papers on the topic of ``Computer Vision in the Wild (CVinW)''

GeoChat. [CVPR 2024 🔥] GeoChat, the first grounded Large Vision Language Model for Remote Sensing

DeepOSM. Train a deep learning net with OpenStreetMap features and satellite imagery.

SSP. Our Implement for SSP

DINOv. [CVPR 2024] Official implementation of the paper "Visual In-context Learning"

DiffNCuts. This is the implementation of our ECCV 2024 paper "Unsupervised Dense Prediction using Differentiable Normalized Cuts" by Yanbin Liu and Stephen Gould.

GraphAdapter. The efficient tuning method for VLMs

SUMformer. Rethinking Urban Mobility Prediction: A Super-Multivariate Time Series Forecasting Approach (TITS)

UniST. Official implementation for "UniST: A Prompt-Empowered Universal Model for Urban Spatio-Temporal Prediction" (KDD 2024)

lmms-eval. One-for-All Multimodal Evaluation Toolkit Across Text, Image, Video, and Audio Tasks

ssl_eval_protocols. Code for the paper "A Closer Look at Benchmarking Self-Supervised Pre-training with Image Classification"

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

SlowFast. PySlowFast: video understanding codebase from FAIR for reproducing state-of-the-art video models.

AI_for_Science_paper_collection. List the AI for Science papers accepted by top conferences

Multi-Level-Logit-Distillation. Code for 'Multi-level Logit Distillation' (CVPR2023)

atlas. [NeurIPS'24] Official PyTorch implementation for paper "Knowledge Composition using Task Vectors with Learned Anisotropic Scaling"

RepDistiller. [ICLR 2020] Contrastive Representation Distillation (CRD), and benchmark of recent knowledge distillation methods

NOAH. [TPAMI] Searching prompt modules for parameter-efficient transfer learning.

leopart. Python

low_cost_robot. Python

action_seg_ot. [CVPR 2024] - Official code for the paper "Temporally Consistent Unbalanced Optimal Transport for Unsupervised Action Segmentation"

mamba. Mamba SSM architecture

chronos-forecasting. Chronos: Pretrained Models for Time Series Forecasting

demistfy_correspondence. Code for the ECCV22 paper Demystifying Unsupervised Semantic Correspondence Estimation

Candidate-Reranking-CIR. The official implementation for Candidate Set Re-ranking for Composed Image Retrieval (TMLR) 01/2024

Neural-Network-Diffusion. We introduce a novel approach for parameter generation, named neural network parameter diffusion (p-diff), which employs a standard latent diffusion model to synthesize a new set of parameters

POT. POT : Python Optimal Transport

VLGuard. [ICML 2024] Safety Fine-Tuning at (Almost) No Cost: A Baseline for Vision Large Language Models.

Depth-Anything. [CVPR 2024] Depth Anything: Unleashing the Power of Large-Scale Unlabeled Data. Foundation Model for Monocular Depth Estimation

snntorch. Deep and online learning with spiking neural networks in Python

Monocular-Depth-Estimation-Toolbox. Monocular Depth Estimation Toolbox based on MMSegmentation.

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

Q-Instruct. ②[CVPR 2024] Low-level visual instruction tuning, with a 200K dataset and a model zoo for fine-tuned checkpoints.

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

Amodal-Completion-in-the-Wild. Python

FusedGW-Entity-Alignment. "A Fused Gromov-Wasserstein Framework for Unsupervised Knowledge Graph Entity Alignment" in ACL 2023

LLM-Shearing. [ICLR 2024] Sheared LLaMA: Accelerating Language Model Pre-training via Structured Pruning

DejaVu. Python

weight-selection. Python

Awesome-Pruning. A curated list of neural network pruning resources.

meta-omnium. Implementation of "Meta Omnium: A Benchmark for General-Purpose Learning-to-Learn"

OneLLM. [CVPR 2024] OneLLM: One Framework to Align All Modalities with Language

yet-another-gpt-tutorial-v2. Jupyter Notebook

objaverse-xl. 🪐 Objaverse-XL is a Universe of 10M+ 3D Objects. Contains API Scripts for Downloading and Processing!

3dv_tutorial. An Invitation to 3D Vision: A Tutorial for Everyone

open_clip. An open source implementation of CLIP.

ImageBind. ImageBind One Embedding Space to Bind Them All

LLMsPracticalGuide. A curated list of practical guide resources of LLMs (LLMs Tree, Examples, Papers)

FoolyourVLLMs. [ICML 2024] Fool Your (Vision and) Language Model With Embarrassingly Simple Permutations

guijiejie.github.io. HTML

dl_tutorials_10weeks. Deep Learning Tutorials for 10 Weeks

intro-to-linear-algebra. Jupyter Notebook

rl_tutorial. Yet Another Reinforcement Learning Tutorial

yet-another-gpt-tutorial. Jupyter Notebook

yet-another-rl-tutorial. Jupyter Notebook

cs231n.github.io. Public facing notes page

SyntheticTumors. [CVPR 2023] Label-Free Liver Tumor Segmentation

Contrastive-Correspondence. Codes for "Learning Contrastive Representation for Semantic Correspondence"

SimpleCRF. matlab and python wrap of crf and dense crf, both 2d and 3d are supported

PyTorch-Gumbel-Sigmoid. Implementation of the Gumbel-Sigmoid distribution in PyTorch.

pytorch-gradual-warmup-lr. Gradually-Warmup Learning Rate Scheduler for PyTorch

TokenCut. (CVPR 2022) Pytorch implementation of "Self-supervised transformers for unsupervised object discovery using normalized cut"

fiftyone. Refine high-quality datasets and visual AI models

awesome-tips.

CutLER. Code release for "Cut and Learn for Unsupervised Object Detection and Instance Segmentation" and "VideoCutLER: Surprisingly Simple Unsupervised Video Instance Segmentation"

Segment-and-Track-Anything. An open-source project dedicated to tracking and segmenting any objects in videos, either automatically or interactively. The primary algorithms utilized include the Segment Anything Model (SAM) for key-frame segmentation and Associating Objects with Transformers (AOT) for efficient tracking and propagation purposes.

TimeCycle. Learning Correspondence from the Cycle-consistency of Time (CVPR 2019)

OpenSeeD. [ICCV 2023] Official implementation of the paper "A Simple Framework for Open-Vocabulary Segmentation and Detection"

codon. A high-performance, zero-overhead, extensible Python compiler with built-in NumPy support

deep-spectral-segmentation. [CVPR 2022] Deep Spectral Methods: A Surprisingly Strong Baseline for Unsupervised Semantic Segmentation and Localization

transformer. Transformer: PyTorch Implementation of "Attention Is All You Need"

TRASHED2022-how-to-record-lectures. Python

learned_optimization. Python

einops. Flexible and powerful tensor operations for readable and reliable code (for pytorch, jax, TF and others)

stable-diffusion. A latent text-to-image diffusion model

nerf-pytorch. A PyTorch re-implementation of Neural Radiance Fields

giraffe. This repository contains the code for the CVPR 2021 paper "GIRAFFE: Representing Scenes as Compositional Generative Neural Feature Fields"

editnerf. Editing a Conditional Radiance Field

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

Bag_of_Tricks_for_Image_Classification_with_Convolutional_Neural_Networks. experiments on Paper <Bag of Tricks for Image Classification with Convolutional Neural Networks> and other useful tricks to improve CNN acc

nerf-pytorch. A PyTorch implementation of NeRF (Neural Radiance Fields) that reproduces the results.

vit-pytorch. Implementation of Vision Transformer, a simple way to achieve SOTA in vision classification with only a single transformer encoder, in Pytorch

ToMe. A method to increase the speed and lower the memory footprint of existing vision transformers.

clrs. Jupyter Notebook

Awesome-Multi-label-Image-Recognition. Awesome Multi-label Image Recognition Paper List

cvpr20_LESA. Shared Attention for Multi-label Zero-shot Learning accepted @ CVPR20

LaSO. LaSO: Label-Set Operations networks for multi-label few-shot learning - official implementation

FewShotMultiLabel. Code for AAAI2021 paper: Few-Shot Learning for Multi-label Intent Detection.

optnet. OptNet: Differentiable Optimization as a Layer in Neural Networks

jaxopt. Hardware accelerated, batchable and differentiable optimizers in JAX.

3D-Medical-Generative-Survey.

CLIP. CLIP (Contrastive Language-Image Pretraining), Predict the most relevant text snippet given an image

treeOT. Python

learning_minimal. Learning to Solve Hard Minimal Problems

VS-Survey.

jaxgptoolbox. Geometry processing utilities compatible with jax for autodifferentiation.

FractalDB-Pretrained-ResNet-PyTorch. Pre-training without Natural Images (ACCV 2020 Best Paper Honorable Mention Award)

FROT. Feature Robust Optimal Transport

neurite. Neural networks toolbox focused on medical image analysis

Sharp-MAML. Sharp-MAML: Sharpness-Aware Model-Agnostic Meta Learning

AdaContrast. Official repository of CVPR '22 paper "Contrastive Test-Time Adaptation"

head-pulse-track. Detecting Pulse from Head Motions in Video

Latex-Template-Rice-USTC. A Latex template including book, thesis, and assignment for Rice and USTC university.

ddn. Deep Declarative Networks

Efficient-3DCNNs. PyTorch Implementation of "Resource Efficient 3D Convolutional Neural Networks", codes and pretrained models.

HowToCook. Programmer's guide about how to cook at home.

LASEM. Code for the ICML 2021 paper "Sharing Less is More: Lifelong Learning in Deep Networks with Selective Layer Transfer"

ott. Optimal transport tools implemented with the JAX framework, to solve large scale matching problems of any flavor.

pytorch-meta-dataset. A non-official 100% PyTorch implementation of META-DATASET benchmark for few-shot classification

generalized-category-discovery. Code for our CVPR 2022 paper 'Generalized Category Discovery'. Project page: https://www.robots.ox.ac.uk/~vgg/research/gcd/

Few_Shot_Distribution_Calibration. [ICLR2021 Oral] Free Lunch for Few-Shot Learning: Distribution Calibration

MixCo-Mixup-Contrast. (NeurIPS 2020 Workshop on SSL) Official Implementation of "MixCo: Mix-up Contrastive Learning for Visual Representation"

tri-M-ICCV. A multi-mode modulator for multi-domain few-shot classification (ICCV)

eg3d. Python

stylegan3. Official PyTorch implementation of StyleGAN3

MedicalNet. Many studies have shown that the performance on deep learning is significantly affected by volume of training data. The MedicalNet project provides a series of 3D-ResNet pre-trained models and relative code.

MedicalZooPytorch. A pytorch-based deep learning framework for multi-modal 2D/3D medical image segmentation

ACSConv. [IEEE JBHI'21] Reinventing 2D Convolutions for 3D Images - 1 line of code to convert pretrained 2D models to 3D!

pytorch-fid. Compute FID scores with PyTorch.

simple-pytorch-3dgan. A simple and unofficial 3D-GAN implementation using PyTorch [NeurIPS 2016]

3D-GAN-pytorch. A pytorch implementation of 3D-GAN

3dbraingen. Official Pytorch Implementation of "Generation of 3D Brain MRI Using Auto-Encoding Generative Adversarial Network" (accepted by MICCAI 2019)

medical-datasets. tracking medical datasets, with a focus on medical imaging

MedMNIST. [pip install medmnist] 18x Standardized Datasets for 2D and 3D Biomedical Image Classification

LiftedGAN. (CVPR 2021) Lifting 2D StyleGAN for 3D-Aware Face Generation

3DStyleGAN. 3D StyleGAN2 for Medical Images

cST-ML. The codes and data of paper "cST-ML: Continuous Spatial-Temporal Meta-Learning for Traffic Dynamics Prediction"

pytorch-CycleGAN-and-pix2pix. Image-to-Image Translation in PyTorch

Reg-GAN. Python

Domain-Consensus-Clustering. [CVPR2021] Domain Consensus Clustering for Universal Domain Adaptation

Solacex.