Flash-Attention-Practice An minimal CUDA implementation of FlashAttention v1 and v2 (forward) Usage Please check the corresponding folder