apps/FastDeploy
1
0
Fork 0
mirror of https://github.com/PaddlePaddle/FastDeploy.git synced 2025-10-05 08:37:06 +08:00
Code Issues Actions 2 Packages Projects Releases Wiki Activity

remove cum_offsets from get_block_shape_and_split_kv_block (#2913)

Browse Source 
* remove padding_offsets from get_padding_offset.cu

* remove padding_offsets from get_padding_offset.cu

* remove padding_offsets from get_padding_offset.cu

* remove cum_offsets from get_block_shape_and_split_kv_block

* remove cum_offsets from get_block_shape_and_split_kv_block
This commit is contained in:
周周周
2025-07-18 16:13:32 +08:00
committed by GitHub
parent e81137e581
commit d306944f4f
6 changed files with 6 additions and 14 deletions
Download Patch File Download Diff File Expand all files Collapse all files

1
fastdeploy/model_executor/layers/attention/flash_attn_backend.py
View File

@@ -151,7 +151,6 @@ class FlashAttentionBackend(AttentionBackend):
forward_meta.seq_lens_encoder,
forward_meta.seq_lens_decoder,
forward_meta.seq_lens_this_time,
forward_meta.cum_offsets,
metadata.encoder_block_shape_q,
metadata.decoder_block_shape_q,
self.num_heads // self.kv_num_heads,