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

flashinfer: remove contiguous calls

Browse Source
This commit is contained in:
Daniël de Kok 2025-01-06 16:07:58 +00:00
parent 02e3dc49be
commit 6defe57d7a
1 changed files with 2 additions and 4 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

6
server/text_generation_server/layers/attention/cuda.py
View File

@ -60,8 +60,7 @@ def paged_attention(
from text_generation_server.layers.attention.flashinfer import decode_state
return decode_state.get().forward(
# TODO: remove `contiguous` call once https://github.com/flashinfer-ai/flashinfer/pull/553 is merged.
query.contiguous(),
query,
paged_kv_cache=(kv_cache.key, kv_cache.value),
logits_soft_cap=softcap,
sm_scale=softmax_scale,
@ -231,8 +230,7 @@ def attention(
softcap = 0.0
return prefill_with_paged_kv_state.get().forward(
# TODO: remove `contiguous` call once https://github.com/flashinfer-ai/flashinfer/pull/553 is merged.
query.contiguous(),
query,
causal=causal,
paged_kv_cache=(kv_cache.key, kv_cache.value),
logits_soft_cap=softcap,