Skip to content

[SB] relax constraint on min number of new tokens #322

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
merged 2 commits into from
Jul 21, 2025
Merged

[SB] relax constraint on min number of new tokens #322

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
b104329
relax constraint of min num output tokens from 3 to 2
yannicks1 Jul 17, 2025
243d571
refactor test script with for loop
yannicks1 Jul 17, 2025
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
32 changes: 7 additions & 25 deletions tests/e2e/test_spyre_basic.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -160,32 +160,14 @@ def test_batch_handling(model: str, backend: str, cb: int,

prompts = get_chicken_soup_prompts(4)

sampling_params1 = SamplingParams(max_tokens=5,
min_tokens=5,
temperature=0,
ignore_eos=True,
logprobs=0)
sampling_params2 = SamplingParams(max_tokens=20,
min_tokens=20,
temperature=0,
ignore_eos=True,
logprobs=0)
sampling_params3 = SamplingParams(max_tokens=10,
min_tokens=10,
temperature=0,
ignore_eos=True,
logprobs=0)
sampling_params4 = SamplingParams(max_tokens=5,
min_tokens=5,
temperature=0,
ignore_eos=True,
logprobs=0)
max_new_tokens = [5, 20, 10, 5]

vllm_sampling_params = [
sampling_params1,
sampling_params2,
sampling_params3,
sampling_params4,
SamplingParams(max_tokens=max_new_tokens[i],
min_tokens=max_new_tokens[i],
temperature=0,
ignore_eos=True,
logprobs=0) for i in range(len(max_new_tokens))
]

kwargs = {
Expand All @@ -207,7 +189,7 @@ def test_batch_handling(model: str, backend: str, cb: int,
**kwargs)
hf_results = generate_hf_output(model=model,
prompts=prompts,
max_new_tokens=[5, 20, 10, 5])
max_new_tokens=max_new_tokens)

compare_results(
model=model,
Expand Down
4 changes: 2 additions & 2 deletions vllm_spyre/v1/worker/spyre_worker.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -79,9 +79,9 @@ def compile_or_warm_up_model(self) -> None:
if self.model_config.task != "embed":
# TODO: remove if spyre supports
# lower number of output tokens
assert num_decode_tokens >= 3, (
assert num_decode_tokens >= 2, (
"VLLM_SPYRE_WARMUP_NEW_TOKENS must be "
"at least 3 (spyre requirement).")
"at least 2 (spyre requirement).")
# warmup individual combination
logger.info(
"[WARMUP] (%d/%d) for prompt length %d, decoding %d tokens "
Expand Down
4 changes: 2 additions & 2 deletions vllm_spyre/worker/spyre_worker.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -208,9 +208,9 @@ def load_model(self):
if self.model_config.task != "embed":
# TODO: remove if spyre supports
# lower number of output tokens
assert num_decode_tokens >= 3, (
assert num_decode_tokens >= 2, (
"VLLM_SPYRE_WARMUP_NEW_TOKENS must be "
"at least 3 (spyre requirement).")
"at least 2 (spyre requirement).")
# warmup individual combination
print(f"[SpyreWorker] Warmup {i+1}/"
f"{len(wup_new_tokens)} "
Expand Down
Loading