Upload generated eval data for npr (with 'eval_size' of 320)

EVAL DATASET GENERATED FROM ORIGINAL DATASET'S TRAINING SPLIT (with 'eval_size' of 320)
Hard negative mining performed AFTER train/eval splitting original pool size: 377840; so, this eval split may contain negatives from training split)

Source dataset: sentence-transformers/npr
(dataset name: npr).
Hard negatives mined with model: sentence-transformers/all-MiniLM-L6-v2.
Mining params: max_score=0.8, num_negatives=1, relative_margin=0.05, absolute_margin=0.03, range_max=100.

README.md

@@ -46,8 +46,11 @@ dataset_info:
   - name: train
     num_bytes: 255693816
     num_examples: 268879
-  download_size: 172423419
-  dataset_size: 255693816
+  - name: eval
+    num_bytes: 285734
+    num_examples: 320
+  download_size: 172624074
+  dataset_size: 255979550
 - config_name: paws-1HN
   features:
   - name: sentence1
@@ -133,6 +136,8 @@ configs:
   data_files:
   - split: train
     path: npr-1HN/train-*
+  - split: eval
+    path: npr-1HN/eval-*
 - config_name: paws-1HN
   data_files:
   - split: train

npr-1HN/eval-00000-of-00001.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:249ade49c5ba208cf2aaa9d486d42c006f537fac519a503aec1202d5f3d437db
+size 200655