Skip to content

fix(judge): honor concurrent_requests for the OpenAI/TGI judge backend #1371

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
mohansree14 wants to merge 1 commit into
base: main
Choose a base branch
from
+4 −2
Open

fix(judge): honor concurrent_requests for the OpenAI/TGI judge backend #1371

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
6 changes: 4 additions & 2 deletions src/lighteval/metrics/utils/llm_as_judge.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -80,7 +80,8 @@ class JudgeLM:
hf_provider (Literal["black-forest-labs", "cerebras", "cohere", "fal-ai", "fireworks-ai",
"inference-providers", "hyperbolic", "nebius", "novita", "openai", "replicate", "sambanova", "together"] | None):
The HuggingFace provider when using the inference-providers backend.
backend_options (dict | None): Options for the backend. Currently only supported for litellm.
backend_options (dict | None): Options for the backend. The "concurrent_requests" key (default 10)
is honored by every backend; the rest of the options are currently only supported for litellm.

Methods:
evaluate_answer: Evaluates an answer using the OpenAI API or Transformers library.
Expand Down Expand Up @@ -416,7 +417,8 @@ async def __call_hf_inference(self, prompt):

def __call_api_parallel(self, prompts):
results = []
with ThreadPoolExecutor(10) as executor:
concurrent_requests = self.backend_options.get("concurrent_requests", 10)
with ThreadPoolExecutor(concurrent_requests) as executor:
for entry in tqdm(executor.map(self.__call_api, prompts), total=len(prompts)):
results.append(entry)

Expand Down