Skip to content

fix: Prevent OverflowError in Offline client #167

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

Open
TheodoreEhrenborg wants to merge 4 commits into
base: main
Choose a base branch
from
Open

fix: Prevent OverflowError in Offline client #167

Changes from all commits
Commits
Show all changes
4 commits
Select commit Hold shift + click to select a range
db089d7
Don't reuse sampling params
TheodoreEhrenborg Dec 19, 2025
7c9bac8
Claude: Use constructor
TheodoreEhrenborg Dec 19, 2025
e0e4646
Shorten line
TheodoreEhrenborg Dec 20, 2025
5459a22
Be less confident in comment
TheodoreEhrenborg Dec 20, 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
24 changes: 18 additions & 6 deletions delphi/clients/offline.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -86,16 +86,28 @@ async def process_func(
Process a single request.
"""

# This is actually stupid
# Extract params from kwargs - must pass to constructor, not mutate after,
# because SamplingParams.__post_init__ likely does some extra setup,
# and mutation after construction skips this.
logprobs = None
prompt_logprobs = None
max_tokens = self.sampling_params.max_tokens
temperature = 1.0
Comment on lines +92 to +95
Copy link
Author

@TheodoreEhrenborg TheodoreEhrenborg Dec 20, 2025

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Not sure if these should default to the default values in the definition of SamplingParams, or default to self.sampling_params's most recent values

for kwarg in kwargs:
if "logprobs" in kwarg:
self.sampling_params.logprobs = kwarg["top_logprobs"]
logprobs = kwarg["top_logprobs"]
if "prompt_logprobs" in kwarg:
self.sampling_params.prompt_logprobs = kwarg["prompt_logprobs"]
prompt_logprobs = kwarg["prompt_logprobs"]
if "max_tokens" in kwarg:
self.sampling_params.max_tokens = kwarg["max_tokens"]
max_tokens = kwarg["max_tokens"]
if "temperature" in kwarg:
self.sampling_params.temperature = kwarg["temperature"]
temperature = kwarg["temperature"]
sampling_params = SamplingParams(
max_tokens=max_tokens,
logprobs=logprobs,
prompt_logprobs=prompt_logprobs,
temperature=temperature,
)
loop = asyncio.get_running_loop()
prompts = []
statistics = []
Expand Down Expand Up @@ -124,7 +136,7 @@ async def process_func(
partial(
self.client.generate, # type: ignore
prompts,
sampling_params=self.sampling_params,
sampling_params=sampling_params, # Use fresh sampling_params
use_tqdm=False,
),
)
Expand Down