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

fix: avoid setting use_sgmv if no kernels present

Browse Source
This commit is contained in:
drbh 2024-12-02 19:40:49 -05:00
parent b57f370386
commit 28ba5e9618
1 changed files with 4 additions and 2 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

6
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,
)
@ -325,8 +326,10 @@ class BatchLoraWeights(BatchAdapterWeights):
default=0,
)
use_sgmv = False
if prefill or max_rank > BGMV_MAX_RANK:
use_sgmv = True
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(
[
(