Ultra-fast dataset upload with pre-encoded bytes

README.md

@@ -102,14 +102,14 @@ dataset_info:
   - name: total
     num_bytes: 896380909
     num_examples: 19880
-  - name: train
-    num_bytes: 691491952
-    num_examples: 15296
   - name: val
     num_bytes: 43548223
     num_examples: 952
-  download_size: 9487350247
-  dataset_size: 1631421084
+  - name: train
+    num_bytes: 860964071
+    num_examples: 19104
+  download_size: 9898515826
+  dataset_size: 1800893203
 configs:
 - config_name: comparison
   data_files:

streaming/train-00000-of-00002.parquet

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:0d078a85531f68f0ed12a900bed480ec733aac57afa49828bc36ea30b9fd2d41
-size 164172876
+oid sha256:10902100ef5c66789be6554ae8c5c12fc47e2f4998edc01ccd983ff7eb5527a1
+size 201213245

streaming/train-00001-of-00002.parquet

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:6ba8d7ab707069b39289dd8811ac31ecdd16e1b5c9446a2aa0446bd7e85306fb
-size 170612201
+oid sha256:a96537b935b1e416b66516c6cf193554e2c85fe981394294f8a20eff840e1032
+size 209952334