From 63ffd1b61368d924b83af5a31c3240f23f6d78dc Mon Sep 17 00:00:00 2001 From: sora <2075279110@qq.com> Date: Fri, 11 Sep 2026 07:20:02 +0000 Subject: [PATCH] Fix NameError _fmt (module-level helper restored); aime/hmmt max_tokens 32768->8192 (non-stream path, gateway honors nothink) Co-Authored-By: Claude --- evalharness/data/datasets/aime24.py | 2 +- evalharness/data/datasets/aime25.py | 2 +- evalharness/data/datasets/aime26.py | 2 +- evalharness/data/datasets/hmmt26.py | 2 +- evalharness/progress/rich_terminal.py | 6 ++++++ 5 files changed, 10 insertions(+), 4 deletions(-) diff --git a/evalharness/data/datasets/aime24.py b/evalharness/data/datasets/aime24.py index abcc067..2361725 100644 --- a/evalharness/data/datasets/aime24.py +++ b/evalharness/data/datasets/aime24.py @@ -13,7 +13,7 @@ from ..spec import DatasetSpec name='aime24', source='HuggingFaceH4/aime_2024', # https://huggingface.co/datasets/HuggingFaceH4/aime_2024 split='train', # the dataset ships a single split - gen_config={'temperature': 1.0, 'max_tokens': 32768}, + gen_config={'temperature': 1.0, 'max_tokens': 8192}, prompt_style='aime_es', # es MathArena template: instruction-first, task_type='math', tags=['math', 'competition'], diff --git a/evalharness/data/datasets/aime25.py b/evalharness/data/datasets/aime25.py index 895d548..ee71103 100644 --- a/evalharness/data/datasets/aime25.py +++ b/evalharness/data/datasets/aime25.py @@ -10,7 +10,7 @@ from ..spec import DatasetSpec name='aime25', source='yentinglin/aime_2025', # https://huggingface.co/datasets/yentinglin/aime_2025 split='train', - gen_config={'temperature': 1.0, 'max_tokens': 32768}, + gen_config={'temperature': 1.0, 'max_tokens': 8192}, prompt_style='aime_es', # es MathArena template: instruction-first, task_type='math', tags=['math', 'competition'], diff --git a/evalharness/data/datasets/aime26.py b/evalharness/data/datasets/aime26.py index fc24abc..377a3b6 100644 --- a/evalharness/data/datasets/aime26.py +++ b/evalharness/data/datasets/aime26.py @@ -10,7 +10,7 @@ from ..spec import DatasetSpec name='aime26', source='MathArena/aime_2026', # curated by MathArena (HuggingFace) split='train', # the dataset ships a single split - gen_config={'temperature': 1.0, 'max_tokens': 32768}, + gen_config={'temperature': 1.0, 'max_tokens': 8192}, prompt_style='aime_es', # es MathArena template: instruction-first, task_type='math', tags=['math', 'competition'], diff --git a/evalharness/data/datasets/hmmt26.py b/evalharness/data/datasets/hmmt26.py index 58402b0..54e01d3 100644 --- a/evalharness/data/datasets/hmmt26.py +++ b/evalharness/data/datasets/hmmt26.py @@ -10,7 +10,7 @@ from ..spec import DatasetSpec name='hmmt26', source='MathArena/hmmt_feb_2026', # curated by MathArena (HuggingFace) split='train', # the dataset ships a single split - gen_config={'temperature': 1.0, 'max_tokens': 32768}, + gen_config={'temperature': 1.0, 'max_tokens': 8192}, prompt_style='imo_es', # es template: Problem: prefix + boxed suffix, task_type='math', tags=['math', 'competition'], diff --git a/evalharness/progress/rich_terminal.py b/evalharness/progress/rich_terminal.py index df5075b..add51fe 100644 --- a/evalharness/progress/rich_terminal.py +++ b/evalharness/progress/rich_terminal.py @@ -13,6 +13,12 @@ from rich.progress import ( ) + +def _fmt(sec): + sec = int(sec) + return f'{sec // 60}m{sec % 60:02d}s' if sec >= 60 else f'{sec}s' + + class RichTerminalProgress: def __init__(self, console=None): # accept an EXTERNAL console: CLI phase messages and the live bar must