Convert dataset to Parquet

Convert dataset to Parquet.

README.md

@@ -36,13 +36,21 @@ dataset_info:
           '25': Z
   splits:
   - name: train
-    num_bytes: 52628800
+    num_bytes: 52114000.0
     num_examples: 124800
   - name: test
-    num_bytes: 8775753
+    num_bytes: 8689953.0
     num_examples: 20800
-  download_size: 36381774
-  dataset_size: 61404553
+  download_size: 58971617
+  dataset_size: 60803953.0
+configs:
+- config_name: emnist-letters
+  data_files:
+  - split: train
+    path: emnist-letters/train-*
+  - split: test
+    path: emnist-letters/test-*
+  default: true
 ---
 
 # Dataset Card for "emnist-letters"

emnist-letters/test-00000-of-00001.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:76421442e28d4ba81c10016df2713d18105df34169f405e0eb43470e88b5d98e
+size 8407069

emnist-letters/train-00000-of-00001.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:756ccd59b40c47fd5a7fc622ed1c34d32576d64edc37aaf9accc89de2bf0bb35
+size 50564548