Rename headers during data loading

src/display/utils.py

@@ -47,30 +47,52 @@ auto_eval_column_dict.append(["revision", ColumnContent, ColumnContent("Model sh
 auto_eval_column_dict.append(["dummy", ColumnContent, ColumnContent("model_name_for_query", "str", False, dummy=True)])
 """
 
-auto_eval_column_dict.append(["eval_name", ColumnContent, ColumnContent("eval_name", "str", True, never_hidden=True)])
-auto_eval_column_dict.append(["precision", ColumnContent, ColumnContent("precision", "str", True)])
-auto_eval_column_dict.append(["hf_model_id", ColumnContent, ColumnContent("hf_model_id", "str", True)])
-auto_eval_column_dict.append(["agree_cs", ColumnContent, ColumnContent("agree_cs", "number", True)])
-auto_eval_column_dict.append(["anli_cs", ColumnContent, ColumnContent("anli_cs", "number", True)])
-auto_eval_column_dict.append(["arc_challenge_cs", ColumnContent, ColumnContent("arc_challenge_cs", "number", True)])
-auto_eval_column_dict.append(["arc_easy_cs", ColumnContent, ColumnContent("arc_easy_cs", "number", True)])
-auto_eval_column_dict.append(["belebele_cs", ColumnContent, ColumnContent("belebele_cs", "number", True)])
-auto_eval_column_dict.append(["ctkfacts_cs", ColumnContent, ColumnContent("ctkfacts_cs", "number", True)])
-auto_eval_column_dict.append(["czechnews_cs", ColumnContent, ColumnContent("czechnews_cs", "number", True)])
-auto_eval_column_dict.append(["fb_comments_cs", ColumnContent, ColumnContent("fb_comments_cs", "number", True)])
-auto_eval_column_dict.append(["gsm8k_cs", ColumnContent, ColumnContent("gsm8k_cs", "number", True)])
-auto_eval_column_dict.append(["klokanek_cs", ColumnContent, ColumnContent("klokanek_cs", "number", True)])
-auto_eval_column_dict.append(["mall_reviews_cs", ColumnContent, ColumnContent("mall_reviews_cs", "number", True)])
-auto_eval_column_dict.append(["mmlu_cs", ColumnContent, ColumnContent("mmlu_cs", "number", True)])
-auto_eval_column_dict.append(["sqad_cs", ColumnContent, ColumnContent("sqad_cs", "number", True)])
-auto_eval_column_dict.append(["subjectivity_cs", ColumnContent, ColumnContent("subjectivity_cs", "number", True)])
-auto_eval_column_dict.append(["truthfulqa_cs", ColumnContent, ColumnContent("truthfulqa_cs", "number", True)])
+auto_eval_column_dict.append(["eval_name", ColumnContent, ColumnContent("Model", "str", True, never_hidden=True)])
+auto_eval_column_dict.append(["precision", ColumnContent, ColumnContent("Precision", "str", True)])
+auto_eval_column_dict.append(["hf_model_id", ColumnContent, ColumnContent("Model link (temporary)", "str", True)])
+auto_eval_column_dict.append(["agree_cs", ColumnContent, ColumnContent("AGREE", "number", True)])
+auto_eval_column_dict.append(["anli_cs", ColumnContent, ColumnContent("ANLI", "number", True)])
+auto_eval_column_dict.append(["arc_challenge_cs", ColumnContent, ColumnContent("ARC-Challenge", "number", True)])
+auto_eval_column_dict.append(["arc_easy_cs", ColumnContent, ColumnContent("ARC-Easy", "number", True)])
+auto_eval_column_dict.append(["belebele_cs", ColumnContent, ColumnContent("Belebele", "number", True)])
+auto_eval_column_dict.append(["ctkfacts_cs", ColumnContent, ColumnContent("CTKFacts", "number", True)])
+auto_eval_column_dict.append(["czechnews_cs", ColumnContent, ColumnContent("Czech News", "number", True)])
+auto_eval_column_dict.append(["fb_comments_cs", ColumnContent, ColumnContent("Facebook Comments", "number", True)])
+auto_eval_column_dict.append(["gsm8k_cs", ColumnContent, ColumnContent("GSM8K", "number", True)])
+auto_eval_column_dict.append(["klokanek_cs", ColumnContent, ColumnContent("Klokanek", "number", True)])
+auto_eval_column_dict.append(["mall_reviews_cs", ColumnContent, ColumnContent("Mall Reviews", "number", True)])
+auto_eval_column_dict.append(["mmlu_cs", ColumnContent, ColumnContent("MMLU", "number", True)])
+auto_eval_column_dict.append(["sqad_cs", ColumnContent, ColumnContent("SQAD", "number", True)])
+auto_eval_column_dict.append(["subjectivity_cs", ColumnContent, ColumnContent("Subjectivity", "number", True)])
+auto_eval_column_dict.append(["truthfulqa_cs", ColumnContent, ColumnContent("TruthfulQA", "number", True)])
 
 
 # We use make dataclass to dynamically fill the scores from Tasks
 AutoEvalColumn = make_dataclass("AutoEvalColumn", auto_eval_column_dict, frozen=True)
 
 
+HEADER_MAP = {
+    "eval_name": "Model",
+    "precision": "Precision",
+    "hf_model_id": "Model link (temporary)",
+    "agree_cs": "AGREE",
+    "anli_cs": "ANLI",
+    "arc_challenge_cs": "ARC-Challenge",
+    "arc_easy_cs": "ARC-Easy",
+    "belebele_cs": "Belebele",
+    "ctkfacts_cs": "CTKFacts",
+    "czechnews_cs": "Czech News",
+    "fb_comments_cs": "Facebook Comments",
+    "gsm8k_cs": "GSM8K",
+    "klokanek_cs": "Klokanek",
+    "mall_reviews_cs": "Mall Reviews",
+    "mmlu_cs": "MMLU",
+    "sqad_cs": "SQAD",
+    "subjectivity_cs": "Subjectivity",
+    "truthfulqa_cs": "TruthfulQA",
+}
+
+
 ## For the queue columns in the submission tab
 @dataclass(frozen=True)
 class EvalQueueColumn:  # Queue column

src/populate.py

@@ -4,7 +4,7 @@ import numpy as np
 import pandas as pd
 
 from src.display.formatting import has_no_nan_values, make_clickable_model
-from src.display.utils import AutoEvalColumn, EvalQueueColumn
+from src.display.utils import AutoEvalColumn, EvalQueueColumn, HEADER_MAP
 from src.leaderboard.read_evals import get_raw_eval_results
 
 
@@ -13,6 +13,7 @@ def get_leaderboard_df(results_path: str, requests_path: str, cols: list, benchm
     #all_data_json = [v.to_dict() for v in raw_data]
     df = pd.DataFrame.from_records(raw_data)
     #df = df.sort_values(by=[AutoEvalColumn.average.name], ascending=False)
+    df = df.rename(columns=HEADER_MAP)
     df = df[cols].round(decimals=2)
     df.replace(r'\s+', np.nan, regex=True)
     # filter out if any of the benchmarks have not been produced