huggingface/text-generation-inference
5
0
Fork 0
mirror of https://github.com/huggingface/text-generation-inference.git synced 2025-04-21 23:12:07 +00:00
Code Issues Packages Projects Releases Wiki Activity

fix: avoid setting use_sgmv if no kernels present (#2796)

Browse Source
This commit is contained in:
drbh 2024-12-04 15:26:09 -05:00 committed by GitHub
parent b57f370386
commit e0db633396
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
1 changed files with 4 additions and 2 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

4
server/text_generation_server/adapters/lora.py
View File

@ -24,6 +24,7 @@ from text_generation_server.utils.sgmv import (
orient_for_rank,
pad_rank,
use_cutlass_shrink,
has_sgmv,
)
@ -325,7 +326,9 @@ class BatchLoraWeights(BatchAdapterWeights):
default=0,
)
use_sgmv = False
if prefill or max_rank > BGMV_MAX_RANK:
if has_sgmv():
use_sgmv = True
lora_a_ptr = torch.tensor(
[
@ -352,7 +355,6 @@ class BatchLoraWeights(BatchAdapterWeights):
device=device,
)
else:
use_sgmv = False
lora_a_ptr = torch.tensor(
[
(