Skip to content

[BugFix] qwen2.5vl enable_thinking=true and image_patch_id bug fix #3921

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
Jiang-Jia-Jun merged 1 commit into from
Sep 11, 2025
Merged

[BugFix] qwen2.5vl enable_thinking=true and image_patch_id bug fix #3921

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
12 changes: 11 additions & 1 deletion fastdeploy/input/qwen_vl_processor/qwen_vl_processor.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -69,7 +69,7 @@ def __init__(
tokenizer=self.tokenizer,
**processor_kwargs,
)

self.image_patch_id = self.processor.image_token_id
self.limit_mm_per_prompt = self._parse_limits(limit_mm_per_prompt)

def process_request(self, request, max_model_len=None, **kwargs):
Expand Down Expand Up @@ -249,6 +249,16 @@ def process_request_dict(self, request, max_model_len=None):
# Handle continuation of previous generation by appending existing tokens
if metadata and metadata.get("generated_token_ids"):
self.append_generated_tokens(outputs, metadata["generated_token_ids"])

enable_thinking = False
if metadata:
enable_thinking = metadata.get("enable_thinking", False)

if request.get("chat_template_kwargs"):
chat_template_kwargs = request.get("chat_template_kwargs")
enable_thinking = chat_template_kwargs.get("enable_thinking", False)
request["enable_thinking"] = enable_thinking

outputs = self.pack_outputs(outputs)

request["prompt_token_ids"] = outputs["input_ids"].tolist()
Expand Down
Loading