apps/FastDeploy
1
0
Fork 0
mirror of https://github.com/PaddlePaddle/FastDeploy.git synced 2026-04-23 00:17:25 +08:00
Code Issues Actions 19 Packages Projects Releases Wiki Activity

[XPU] Support CudaGraph(add block attn cuda_graph support) (#6116)

Browse Source 
* add block attn cuda_graph support
This commit is contained in:
yinwei
2026-01-20 19:33:11 +08:00
committed by GitHub
parent 00a6a73431
commit 51a8a2ed57
3 changed files with 76 additions and 11 deletions
Download Patch File Download Diff File Expand all files Collapse all files
fastdeploy/model_executor/layers/backends/xpu/attention.py
+8
View File
@@ -203,6 +203,14 @@ class XPUAttentionBackend(AttentionBackend):
forward_meta.decoder_context_len_cache_cpu,
forward_meta.decoder_batch_map_cpu,
forward_meta.prefix_len_cpu,
forward_meta.encoder_seq_lod,
forward_meta.decoder_seq_lod,
forward_meta.encoder_kv_lod,
forward_meta.encoder_batch_map,
forward_meta.decoder_context_len,
forward_meta.decoder_context_len_cache,
forward_meta.decoder_batch_map,
forward_meta.prefix_len,
cache_k_scale,
cache_v_scale,
cache_k_out_scale,