xinyun/vllm
1
0
Fork 0
mirror of https://git.datalinker.icu/vllm-project/vllm.git synced 2026-01-28 13:27:15 +08:00
Code Issues Packages Projects Releases Wiki Activity

[gpt-oss][Bugfix]Fix streamableparser for missing handling of certain token_ids (#24306)

Browse Source 
Signed-off-by: chaunceyjiang <chaunceyjiang@gmail.com>
This commit is contained in:
Chauncey 2025-09-06 01:26:00 +08:00 committed by GitHub
parent 7812bcf278
commit 23a6c5280e
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
1 changed files with 3 additions and 3 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

6
vllm/entrypoints/context.py
View File

@ -238,11 +238,11 @@ class StreamingHarmonyContext(HarmonyContext):
# (finished=True), then the next token processed will mark the
# beginning of a new message
self.first_tok_of_message = output.finished
tok = output.outputs[0].token_ids[0]
self.parser.process(tok)
for tok in output.outputs[0].token_ids:
self.parser.process(tok)
self._update_num_output_tokens(output.outputs[0].token_ids)
# Check if the current token is part of reasoning content
self._update_num_reasoning_tokens([tok])
self._update_num_reasoning_tokens(output.outputs[0].token_ids)
self.last_tok = tok
else:
# Handle the case of tool output in direct message format