Add FineWeb 10B raw processed dataset with all variants and metadata (120,000 items)

README.md

@@ -0,0 +1,41 @@
+---
+dataset_info:
+  features:
+  - name: variants
+    struct:
+    - name: degenerated
+      dtype: string
+    - name: original
+      dtype: string
+  - name: metadata
+    struct:
+    - name: base
+      dtype: string
+    - name: copy_index
+      dtype: int64
+    - name: degenerate
+      dtype: string
+    - name: duplicate
+      dtype: string
+    - name: is_duplicate_copy
+      dtype: bool
+    - name: is_duplicate_source
+      dtype: bool
+    - name: language
+      dtype: string
+    - name: original_index
+      dtype: int64
+    - name: variety
+      dtype: string
+  splits:
+  - name: train
+    num_bytes: 459357296
+    num_examples: 120000
+  download_size: 292398596
+  dataset_size: 459357296
+configs:
+- config_name: default
+  data_files:
+  - split: train
+    path: data/train-*
+---

data/train-00000-of-00001.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d1801360616612dbf301db2692416e6dac64a2b29d472def05364ef8f6378390
+size 292398596