huggingface/text-generation-inference
5
0
Fork 0
mirror of https://github.com/huggingface/text-generation-inference.git synced 2025-09-11 12:24:53 +00:00
Code Issues Packages Projects Releases Wiki Activity

Fixing the oom maybe with 2.5.1 change.

Browse Source
This commit is contained in:
Nicolas Patry 2025-01-27 14:35:00 +01:00
parent 40b00275b2
commit 57570bf598
No known key found for this signature in database
GPG Key ID: D2920555C90F704C
1 changed files with 1 additions and 1 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
server/text_generation_server/models/globals.py
View File

@ -28,7 +28,7 @@ if PREFIX_CACHING and ATTENTION not in {
raise RuntimeError("Prefix caching is only supported with flashinfer") raise RuntimeError("Prefix caching is only supported with flashinfer")
MEM_POOL = torch.cuda.graph_pool_handle() if torch.cuda.is_available() else None MEM_POOL = torch.cuda.graph_pool_handle() if torch.cuda.is_available() else None
TGI_WIGGLE_ROOM = float(os.getenv("TGI_WIGGLE_ROOM", "0.95")) TGI_WIGGLE_ROOM = float(os.getenv("TGI_WIGGLE_ROOM", "0.93"))
assert TGI_WIGGLE_ROOM > 0 assert TGI_WIGGLE_ROOM > 0
assert TGI_WIGGLE_ROOM < 1 assert TGI_WIGGLE_ROOM < 1