杨奇(yann qi)

R-4B. The official repository of "R-4B: Incentivizing General-Purpose Auto-Thinking Capability in MLLMs via Bi-Mode Integration"

Draw-an-Audio-Code. Official code of the paper: Draw an Audio: Leveraging Multi-Instruction for Video-to-Audio Synthesis.

COMBO-AVS. [CVPR 2024 Highlight] Official implementation of the paper: Cooperation Does Matter: Exploring Multi-Order Bilateral Relations for Audio-Visual Segmentation

RCTS-RAG. [ICML 2025 Spotlight] Official implementation of the paper: Re-ranking Reasoning Context with Tree Search Makes Large Vision-Language Models Stronger

Pixel2Mesh-Tensorflow2. Pixel2Mesh-Tensorflow2

Pytorch-Code-Templete. 用于快速建立一个项目的模板

AVS-COMBO. Official website of the paper: Cooperation Does Matter: Exploring Multi-Order Bilateral Relations for Audio-Visual Segmentation (CVPR 2024)

DyHRNet-in-Remote-Sensing-Images. Dynamic High Resolution Network for Semantic Segmentation in Remote Sensing Images

RefineNet-pytorch. RefineNet by pytorch on COCO Dataset

Draw-an-Audio. Official website of the paper: Draw an Audio: Leveraging Multi-Instruction for Video-to-Audio Synthesis.

SSD-for-pytorch. This repository provides a script and recipe to train the SSD300 model to achieve state of the art accuracy. The codes of model architecture comes from NVIDIA(NVIDIA SSD pytorch), and the method belongs to paper SSD: Single Shot MultiBox Detector .

Bag-of-Tricks-for-CV-Chinese-only. Python

Old_projects_temp_storage. Blog

yannqi.github.io. My main page.

R-4B-VL. Official website of R-4B

CV-Dataset-Dataloader-Pytorch. A CV Dataset template, you can load the dataset conveniently by Pytorch according to this template.

yannqi. Config files for my GitHub profile.

TensorFlow-lite-Android-digital_recogniztion. 基于TensorFlow-lite和Android studio的在安卓手机上实现手写数字识别功能