Skip to content
GitLab
Menu
Projects
Groups
Snippets
Loading...
Help
Help
Support
Community forum
Keyboard shortcuts
?
Submit feedback
Contribute to GitLab
Sign in
Toggle navigation
Menu
Open sidebar
OpenDAS
vllm_cscc
Commits
259605da
"vllm/model_executor/models/commandr.py" did not exist on "37ca5581039271d4bf69e5cb1f1ec8e04775777c"
Commit
259605da
authored
Dec 12, 2025
by
zhuwenwen
Browse files
update fp8_mqa_logits and fp8_paged_mqa_logits
parent
a55b8f91
Changes
1
Show whitespace changes
Inline
Side-by-side
Showing
1 changed file
with
2 additions
and
2 deletions
+2
-2
vllm/model_executor/models/deepseek_v2.py
vllm/model_executor/models/deepseek_v2.py
+2
-2
No files found.
vllm/model_executor/models/deepseek_v2.py
View file @
259605da
...
...
@@ -627,7 +627,7 @@ def sparse_attn_indexer(
chunk
.
cu_seq_lens
,
chunk
.
num_reqs
,
)
if
current_platform
.
is_rocm
():
if
not
current_platform
.
is_rocm
():
logits
=
fp8_mqa_logits
(
q_fp8
[
chunk
.
token_start
:
chunk
.
token_end
],
(
k_fp8
,
k_scale
),
...
...
@@ -680,7 +680,7 @@ def sparse_attn_indexer(
next_n
=
padded_q_fp8_decode_tokens
.
shape
[
1
]
assert
batch_size
==
decode_metadata
.
seq_lens
.
shape
[
0
]
num_padded_tokens
=
batch_size
*
next_n
if
current_platform
.
is_rocm
():
if
not
current_platform
.
is_rocm
():
logits
=
fp8_paged_mqa_logits
(
padded_q_fp8_decode_tokens
,
kv_cache
,
...
...
Write
Preview
Markdown
is supported
0%
Try again
or
attach a new file
.
Attach a file
Cancel
You are about to add
0
people
to the discussion. Proceed with caution.
Finish editing this message first!
Cancel
Please
register
or
sign in
to comment