Skip to content

bench: greedy Qwen3.5 cells and ITL in results.json #909

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
Gaurav-Shah05 merged 2 commits into from
Oct 27, 2026
Merged

bench: greedy Qwen3.5 cells and ITL in results.json #909

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
da469cd
bench: greedy + same think block for qwen3.5, full-precision itl in r…
merceod Oct 27, 2026
e82e4e9
bench: send enable_thinking flat for mstar's qwen3.5
Gaurav-Shah05 Oct 27, 2026
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
7 changes: 7 additions & 0 deletions benchmark/base.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -314,6 +314,13 @@ def get_hf_url(self):
def get_supported_modalities(self):
return {RequestType.T2T, RequestType.I2T, RequestType.V2T}

def get_model_kwargs(self, request_type: RequestType):
# Greedy everywhere (mstar's sampler defaults to 0.6), and the same prompt
# render: mstar opens a <think> block by default, while the 0.8B/2B templates
# only do when asked. mstar reads the flat enable_thinking and vLLM/SGLang
# read chat_template_kwargs.
return {"temperature": 0.0, "enable_thinking": True, "chat_template_kwargs": {"enable_thinking": True}}


class Qwen3_8_Dense(Qwen3_5_Dense):
"""Qwen3.8 dense: the Qwen3.5 architecture, defaulting to the 27B."""
Expand Down
5 changes: 5 additions & 0 deletions benchmark/runner.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -284,6 +284,11 @@ def _stats(stats):
"aggregate": {
"max_concurrency": agg.max_concurrency,
"ttft_s": {modality: _stats(stats) for modality, stats in agg.ttft.items()},
"itl_s": {modality: _stats(stats) for modality, stats in agg.itl.items()},
"text_token_throughput": agg.text_token_throughput,
"total_text_tokens": agg.total_text_tokens,
"mean_text_tokens": agg.mean_text_tokens,
"total_input_tokens": agg.total_input_tokens,
"e2e_latency_s": _stats(agg.e2e_latency),
"rtf": _stats(agg.rtf),
"audio_seconds_throughput": agg.audio_seconds_throughput,
Expand Down
Loading