xinyun/vllm
1
0
Fork 0
mirror of https://git.datalinker.icu/vllm-project/vllm.git synced 2026-06-10 08:49:06 +08:00
Code Issues Packages Projects Releases Wiki Activity

[Misc] Minor code simplification for spec decode (#24053)

Browse Source 
Signed-off-by: Woosuk Kwon <woosuk.kwon@berkeley.edu>
This commit is contained in:
Woosuk Kwon 2025-09-01 08:54:01 -07:00 committed by GitHub
parent 41c80698b3
commit 39a22dcaac
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
1 changed files with 7 additions and 7 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

14
vllm/v1/core/sched/scheduler.py
View File

@ -873,19 +873,19 @@ class Scheduler(SchedulerInterface):
scheduled_spec_token_ids = ( scheduled_spec_token_ids = (
scheduler_output.scheduled_spec_decode_tokens.get(req_id)) scheduler_output.scheduled_spec_decode_tokens.get(req_id))
if scheduled_spec_token_ids: if scheduled_spec_token_ids:
num_draft_tokens = len(scheduled_spec_token_ids)
num_accepted = len(generated_token_ids) - 1
num_rejected = num_draft_tokens - num_accepted
# num_computed_tokens represents the number of tokens # num_computed_tokens represents the number of tokens
# processed in the current step, considering scheduled # processed in the current step, considering scheduled
# tokens and rejections. If some tokens are rejected, # tokens and rejections. If some tokens are rejected,
# num_computed_tokens is decreased by the number of rejected # num_computed_tokens is decreased by the number of rejected
# tokens, where is given by: # tokens.
# len(scheduled_spec_token_ids) + 1 - len(generated_token_ids). request.num_computed_tokens -= num_rejected
num_tokens_rejected = (len(scheduled_spec_token_ids) + 1 -
len(generated_token_ids))
request.num_computed_tokens -= num_tokens_rejected
spec_decoding_stats = self.make_spec_decoding_stats( spec_decoding_stats = self.make_spec_decoding_stats(
spec_decoding_stats, spec_decoding_stats,
num_draft_tokens=len(scheduled_spec_token_ids), num_draft_tokens=num_draft_tokens,
num_accepted_tokens=len(generated_token_ids) - 1) num_accepted_tokens=num_accepted)
stopped = False stopped = False
new_logprobs = None new_logprobs = None