Dear Guo Yejun and Steven Liu, I hope this email finds you well. I am very much interested in the "Modernizing LibTorch Backend: Async Infrastructure, Batching, and GPU Pipelines" project for Google Summer of Code 2026.
I have carefully read the project description, including the goals around migrating to the unified DNNAsyncExecModule, implementing high-throughput batching with at::cat()/at::split(), zero-copy GPU pipeline using at::from_blob() for AV_PIX_FMT_CUDA frames, the Lazy Reallocator, and the FATE test suite. I have 2 years of experience in C++, C systems programming and am currently familiarizing myself with LibTorch (PyTorch C++ API), FFmpeg's libavfilter framework. I would greatly appreciate any guidance on next steps, or any updates to the project scope. More than any other tools out there, ffmpeg has shown me that software has real and meaningful effects on the lives of countless people ,and the best way I can give back is by contributing to this project. Thank you for your time and for mentoring this interesting project. I look forward to your reply. Best regards, t.k.naveen, Chennai / Indian Standard Time _______________________________________________ ffmpeg-devel mailing list -- [email protected] To unsubscribe send an email to [email protected]
