I'm a full-time senoir algorithm engineer. My current work mainly includes on-device LLM & VLM development and accelerating for autonomous vehicles and smart home devices. I am also keen on creating interesting applications using AIGC. Before I joined the industry, I worked on computer vision during my PhD (I was selected as one of the world’s top 2% research scientists), and mobile game development during my Master (I developed two indie IOS games in 2015 and also took a internship at miHoYo in 2017). My lifelong dream is to contribute to the development of a world-changing products that benefit millions of people.
Pinned Loading
-
Scene-Graph-Benchmark.pytorch
Scene-Graph-Benchmark.pytorch PublicA new codebase for popular Scene Graph Generation methods (2020). Visualization & Scene Graph Extraction on custom images/datasets are provided. It's also a PyTorch implementation of paper “Unbiase…
-
Long-Tailed-Recognition.pytorch
Long-Tailed-Recognition.pytorch Public[NeurIPS 2020] This project provides a strong single-stage baseline for Long-Tailed Classification, Detection, and Instance Segmentation (LVIS). It is also a PyTorch implementation of the NeurIPS 2…
-
VQA2.0-Recent-Approachs-2018.pytorch
VQA2.0-Recent-Approachs-2018.pytorch PublicA pytroch reimplementation of "Bilinear Attention Network", "Intra- and Inter-modality Attention", "Learning Conditioned Graph Structures", "Learning to count object", "Bottom-up top-down" for Visu…
-
VCTree-Scene-Graph-Generation
VCTree-Scene-Graph-Generation PublicCode for the Scene Graph Generation part of CVPR 2019 oral paper: "Learning to Compose Dynamic Tree Structures for Visual Contexts"
-
Generalized-Long-Tailed-Benchmarks.pytorch
Generalized-Long-Tailed-Benchmarks.pytorch Public[ECCV 2022] A generalized long-tailed challenge that incorporates both the conventional class-wise imbalance and the overlooked attribute-wise imbalance within each class. The proposed IFL together…
-
Qwen-Tokenizer-Pruner
Qwen-Tokenizer-Pruner PublicDue to the huge vocaburary size (151,936) of Qwen models, the Embedding and LM Head weights are excessively heavy. Therefore, this project provides a Tokenizer vocabulary shearing solution for Qwen…
If the problem persists, check the GitHub status page or contact support.