Save GLOBE_V2 test split as train split for cross validation

README.md

@@ -0,0 +1,29 @@
+---
+dataset_info:
+  features:
+  - name: audio
+    dtype: audio
+  - name: speaker_id
+    dtype: string
+  - name: transcript
+    dtype: string
+  - name: accent
+    dtype: string
+  - name: duration
+    dtype: float64
+  - name: age
+    dtype: string
+  - name: gender
+    dtype: string
+  splits:
+  - name: train
+    num_bytes: 755286998.0
+    num_examples: 5046
+  download_size: 722477750
+  dataset_size: 755286998.0
+configs:
+- config_name: default
+  data_files:
+  - split: train
+    path: data/train-*
+---

data/train-00000-of-00002.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:bf49e3b7d5153f3ebb0c8f6133bf069aa7946a0d8243357029e4e83c77940365
+size 359974375

data/train-00001-of-00002.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:abdffeb3a3dc0ddd5814b03df755ddc24fadb66d0ca2405ea04c539541ac28af
+size 362503375