corpus uses test split; nq-hard uses test split

README.md

@@ -2,7 +2,7 @@
 configs:
 - config_name: ConditionalQA-corpus
   data_files:
-  - split: train
+  - split: test
     path: ConditionalQA/corpus/*
 - config_name: ConditionalQA-queries
   data_files:
@@ -22,11 +22,11 @@ configs:
     path: ConditionalQA/qrels/test.parquet
 - config_name: NaturalQuestions-corpus
   data_files:
-  - split: train
+  - split: test
     path: NaturalQuestions/corpus/*
 - config_name: nq-hard
   data_files:
-  - split: train
+  - split: test
     path: NaturalQuestions/nq-hard/*
 - config_name: NaturalQuestions-queries
   data_files:
@@ -46,7 +46,7 @@ configs:
     path: NaturalQuestions/qrels/test.parquet
 - config_name: Genomics-corpus
   data_files:
-  - split: train
+  - split: test
     path: Genomics/corpus/*
 - config_name: Genomics-queries
   data_files:
@@ -58,7 +58,7 @@ configs:
     path: Genomics/qrels/test.parquet
 - config_name: MSMARCO-corpus
   data_files:
-  - split: train
+  - split: test
     path: MSMARCO/corpus/*
 - config_name: MSMARCO-queries
   data_files:
@@ -78,7 +78,7 @@ configs:
     path: MSMARCO/qrels/test.parquet
 - config_name: MIRACL-corpus
   data_files:
-  - split: train
+  - split: test
     path: MIRACL/corpus/*
 - config_name: MIRACL-queries
   data_files: