xeon27 commited on
Commit
b1f9063
·
1 Parent(s): 116683a
Files changed (1) hide show
  1. src/populate.py +2 -4
src/populate.py CHANGED
@@ -1,6 +1,7 @@
1
  import json
2
  import os
3
 
 
4
  import pandas as pd
5
 
6
  from src.about import Tasks
@@ -44,12 +45,9 @@ def get_leaderboard_df(results_path: str, requests_path: str, cols: list, benchm
44
  # # filter out if any of the benchmarks have not been produced
45
  # df = df[has_no_nan_values(df, benchmark_cols)]
46
 
47
- # TMP: Log
48
- print(df.shape)
49
-
50
  # make values clickable and link to log files
51
  for col in benchmark_cols:
52
- df[col] = df[[AutoEvalColumn.model.name, col]].apply(lambda x: f"[{x[col]}]({get_inspect_log_url(model_name=x[AutoEvalColumn.model.name].split('>')[1].split('<')[0], benchmark_name=TASK_NAME_INVERSE_MAP[col]['name'])})" if x[col] is not None else "-", axis=1)
53
 
54
  # # make task names clickable and link to inspect-evals repository - this creates issues later
55
  # df = df.rename(columns={col: f"[{col}]({TASK_NAME_INVERSE_MAP[col]['source']})" for col in benchmark_cols})
 
1
  import json
2
  import os
3
 
4
+ import numpy as np
5
  import pandas as pd
6
 
7
  from src.about import Tasks
 
45
  # # filter out if any of the benchmarks have not been produced
46
  # df = df[has_no_nan_values(df, benchmark_cols)]
47
 
 
 
 
48
  # make values clickable and link to log files
49
  for col in benchmark_cols:
50
+ df[col] = df[[AutoEvalColumn.model.name, col]].apply(lambda x: f"[{x[col]}]({get_inspect_log_url(model_name=x[AutoEvalColumn.model.name].split('>')[1].split('<')[0], benchmark_name=TASK_NAME_INVERSE_MAP[col]['name'])})" if x[col] != np.nan else "-", axis=1)
51
 
52
  # # make task names clickable and link to inspect-evals repository - this creates issues later
53
  # df = df.rename(columns={col: f"[{col}]({TASK_NAME_INVERSE_MAP[col]['source']})" for col in benchmark_cols})