tspeterkim
- Karma
- 186
- Created
- March 7, 2024 (2y ago)
- Submissions
- 0
https://tspeterkim.github.io/
- I came back to school to study hardware after 5 years of doing ML (tspeterkim.github.io)
- I came back to school to study hardware after 5 years of doing ML (tspeterkim.github.io)
- How to speed up Flux inference by 40% in one line (twitter.com)
- Show HN: Minimal Paged Attention (github.com)
- Show HN: Mixed Precision Training from Scratch (github.com)
Detailed writeup here: https://tspeterkim.github.io/posts/mixed-precision-from-scra...
- Mixed Precision Training from Scratch (tspeterkim.github.io)
- Profiling Remote GPUs with Nsight Compute (tspeterkim.github.io)
- Show HN: One Billion Rows in CUDA (github.com)
- The One Billion Row Challenge in CUDA (tspeterkim.github.io)
- Show HN: Flash Attention in ~100 lines of CUDA (github.com)