An open-source book and code collection of modern CUDA learning notes, with 200+ kernels covering HGEMM, FP8/FP4 GEMM, and Flash-Attention.
12.1K1.3KCudaGPL-3.0Updated 17 hours ago
#flash-attention
Open-source projects tagged “flash-attention”
Curated GitHub repositories related to flash-attention, with AI summaries and key stats.