dearwolf/aider
1
0
Fork 0
mirror of https://github.com/Aider-AI/aider.git synced 2025-05-21 21:04:59 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions 5

refactor: Use constant for hard set threshold

Browse source
This commit is contained in:
Paul Gauthier (aider) 2024-12-18 13:06:03 -08:00
parent 051cabed69
commit e6bfc1c2fc
1 changed files with 6 additions and 3 deletions
Download patch file Download diff file Expand all files Collapse all files

9
benchmark/problem_stats.py
View file

@ -3,6 +3,9 @@
import argparse import argparse
import json import json
from collections import defaultdict from collections import defaultdict
from typing import List, Optional
HARD_SET_NUM = 4 # Number of models that defines the hard set threshold
from pathlib import Path from pathlib import Path
import yaml import yaml
@ -178,10 +181,10 @@ def analyze_exercise_solutions(dirs=None, topn=None):
cumsum += count cumsum += count
print(f"{i:>6d} {count:>9d} {cumsum:>10d}") print(f"{i:>6d} {count:>9d} {cumsum:>10d}")
# Collect the hard set (exercises solved by 4 or fewer models) # Collect the hard set (exercises solved by HARD_SET_NUM or fewer models)
print("\nHard Set Analysis (exercises solved by ≤4 models):") print(f"\nHard Set Analysis (exercises solved by ≤{HARD_SET_NUM} models):")
print("-" * 60) print("-" * 60)
hard_set = {ex for ex, models in exercise_solutions.items() if len(models) <= 4} hard_set = {ex for ex, models in exercise_solutions.items() if len(models) <= HARD_SET_NUM}
print(f"Total hard set exercises: {len(hard_set)}") print(f"Total hard set exercises: {len(hard_set)}")
# For each model, compute performance on hard set # For each model, compute performance on hard set