NousResearch/atropos
1
0
Fork 0
mirror of https://github.com/NousResearch/atropos.git synced 2026-04-19 12:57:58 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions

[pre-commit.ci] auto fixes from pre-commit.com hooks

Browse source 
for more information, see https://pre-commit.ci
This commit is contained in:
pre-commit-ci[bot] 2025-12-28 04:12:13 +00:00
parent ea6db6fe92
commit 1d4275d441
2 changed files with 19 additions and 12 deletions
Download patch file Download diff file Expand all files Collapse all files

24
environments/eval_environments/phybench_eval.py
View file

@ -176,7 +176,6 @@ class PHYBenchEvalConfig(BaseEnvConfig):
)
class PHYBenchEvalEnv(BaseEnv):
"""
PHYBench Evaluation Environment.
@ -496,7 +495,10 @@ class PHYBenchEvalEnv(BaseEnv):
async def rollout_and_score_eval(self, item: Dict) -> Optional[Dict]:
"""Run evaluation on a single item and return the result."""
if self.config.full_debug:
print(f"[DEBUG] Starting eval for item: {item.get('id', 'unknown')}", flush=True)
print(
f"[DEBUG] Starting eval for item: {item.get('id', 'unknown')}",
flush=True,
)
prompt = self._format_prompt(item)
system_content = self._create_system_content()
@ -519,9 +521,17 @@ class PHYBenchEvalEnv(BaseEnv):
for attempt in range(self.config.max_retries):
try:
if self.config.full_debug:
print(f" Making API request (attempt {attempt + 1}/{self.config.max_retries})...", flush=True)
print(f" Temperature: {self.config.eval_temperature}", flush=True)
print(f" Max tokens: {self.config.eval_max_tokens if self.config.eval_max_tokens > 0 else 'model default'}", flush=True)
print(
f" Making API request (attempt {attempt + 1}/{self.config.max_retries})...",
flush=True,
)
print(
f" Temperature: {self.config.eval_temperature}", flush=True
)
print(
f" Max tokens: {self.config.eval_max_tokens if self.config.eval_max_tokens > 0 else 'model default'}",
flush=True,
)
response = await self.server.chat_completion(**kwargs)
response_text = response.choices[0].message.content or ""
@ -610,9 +620,7 @@ class PHYBenchEvalEnv(BaseEnv):
print(f"{'='*60}\n")
# Create evaluation tasks
eval_tasks = [
self.rollout_and_score_eval(item) for item in self.eval_items
]
eval_tasks = [self.rollout_and_score_eval(item) for item in self.eval_items]
# Run with progress bar
results = await tqdm_asyncio.gather(*eval_tasks, desc="Evaluating PHYBench")