Update README.md

README.md

@@ -1,20 +1,5 @@
 ---
 license: mit
-dataset_info:
-  features:
-  - name: tokens
-    dtype: uint16
-    shape: [100000000]
-  config_name: default
-  splits:
-  - name: train
-    num_bytes: 200000000000
-    num_examples: 100
-  - name: val  
-    num_bytes: 200000000
-    num_examples: 1
-  download_size: 20000000000
-  dataset_size: 200200000000
 language:
 - en
 tags:
@@ -29,10 +14,30 @@ tags:
 - numpy
 task_categories:
 - text-generation
-- language-modeling
 pretty_name: FineWeb-Edu 10B Tokens (Tokenized NPY Format)
 size_categories:
 - 10B<n<100B
+dataset_info:
+  features:
+  - name: tokens
+    dtype: uint16
+  config_name: default
+  splits:
+  - name: train
+    num_bytes: 19800000000
+    num_examples: 99
+  - name: val  
+    num_bytes: 200000000
+    num_examples: 1
+  download_size: 20000000000
+  dataset_size: 20000000000
+configs:
+- config_name: default
+  data_files:
+  - split: train
+    path: "edufineweb_train_*.npy"
+  - split: val
+    path: "edufineweb_val_*.npy"
 ---
 
 # FineWeb-Edu 10B Tokens (NPY Format)