huggingface/text-generation-inference
5
0
Fork 0
mirror of https://github.com/huggingface/text-generation-inference.git synced 2025-09-11 04:14:52 +00:00
Code Issues Packages Projects Releases Wiki Activity

fix vllm import error

Browse Source
This commit is contained in:
Zeyu Li 2023-12-30 14:26:37 +08:00
parent 630800eed3
commit ad7f839673
1 changed files with 4 additions and 4 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

8
server/text_generation_server/utils/paged_attention.py
View File

@ -1,8 +1,8 @@
import torch
# vllm imports
from vllm import cache_ops
from vllm import attention_ops
from vllm._C import cache_ops
from vllm._C import ops
_PARTITION_SIZE = 512
@ -56,7 +56,7 @@ def attention(
# to parallelize.
use_v1 = max_num_partitions == 1 or num_seqs * num_heads > 512
if use_v1:
attention_ops.paged_attention_v1(
ops.paged_attention_v1(
out,
query,
key_cache,
@ -83,7 +83,7 @@ def attention(
device=out.device,
)
max_logits = torch.empty_like(exp_sums)
attention_ops.paged_attention_v2(
ops.paged_attention_v2(
out,
exp_sums,
max_logits,