pytorch/benchmarks/transformer
Eddie Yan cd380c794f [CUDNN][SDPA] Experimental cuDNN Flash Attention v2 Inference (#115663)
#113713

Going to clean up some of the checks and will remove draft status after.
Can be tested on SM80+ with `TORCH_CUDNN_MHA_ENABLED=1`.

CC @drisspg @ptrblck
Pull Request resolved: https://github.com/pytorch/pytorch/pull/115663
Approved by: https://github.com/drisspg
2024-02-14 22:02:06 +00:00
..
attention_bias_benchmarks.py [BE]: Apply RUF025 dict.fromkeys preview rule (#118637) 2024-01-30 20:46:54 +00:00
better_transformer_vs_mha_functional.py Apply UFMT to all files in benchmarks/ (#105928) 2023-07-26 01:18:48 +00:00
sdp.py [CUDNN][SDPA] Experimental cuDNN Flash Attention v2 Inference (#115663) 2024-02-14 22:02:06 +00:00
sdpa.py Update the sdpa benchmark to measure forward backward time in isolation (#115986) 2023-12-18 22:40:47 +00:00