xinyun/vllm
1
0
Fork 0
mirror of https://git.datalinker.icu/vllm-project/vllm.git synced 2025-12-20 10:05:54 +08:00
Code Issues Packages Projects Releases Wiki Activity

Enable prefix caching with full cuda graphs (#19617)

Browse Source 
Signed-off-by: Woosuk Kwon <woosuk.kwon@berkeley.edu>
This commit is contained in:
Woosuk Kwon 2025-06-15 01:05:05 -07:00 committed by GitHub
parent 3d330c4c09
commit 055915e6ce
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
1 changed files with 0 additions and 1 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

1
vllm/config.py
View File

@ -4495,7 +4495,6 @@ class VllmConfig:
"full_cuda_graph is not supported with " "full_cuda_graph is not supported with "
"cascade attention. Disabling cascade attention.") "cascade attention. Disabling cascade attention.")
self.model_config.disable_cascade_attn = True self.model_config.disable_cascade_attn = True
self.cache_config.enable_prefix_caching = False
if (self.kv_events_config is not None if (self.kv_events_config is not None
and self.kv_events_config.enable_kv_cache_events and self.kv_events_config.enable_kv_cache_events