Tencent-Hunyuan/FlashTopK
Fast, exact fused top-k selection for learned sparse attention training and inference, from HyScale (Tencent Hy & University of Edinburgh).
GitHub repository with 16 stars and 2 forks.
Language: Python
Topics: cuda, cute-dsl, deepseek-sparse-attention, hyscale, lightning-indexer, llm-inference, llm-training, long-context, pytorch, sparse-attention