Upload README.md with huggingface_hub
Browse files
README.md
ADDED
|
@@ -0,0 +1,22 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
license: mit
|
| 3 |
+
language: en
|
| 4 |
+
tags: [cs336, openwebtext, bpe, tokenized]
|
| 5 |
+
---
|
| 6 |
+
# CS336 OWT 32k BPE tokenized
|
| 7 |
+
|
| 8 |
+
Pre-tokenized OpenWebText sample (Stanford CS336 assignment 1 source) using a custom 32k byte-level BPE tokenizer.
|
| 9 |
+
|
| 10 |
+
## Files
|
| 11 |
+
- `owt_vocab_32k.pkl` — vocab dict\[int, bytes\] (32000 entries; ID 0 = <|endoftext|>, 1..256 = raw bytes, 257..31999 = merges)
|
| 12 |
+
- `owt_merges_32k.pkl` — list\[tuple[bytes, bytes]\] in merge order
|
| 13 |
+
- `owt_train.npy` — uint16 token IDs of 11 GB owt_train.txt (2,729,574,028 tokens)
|
| 14 |
+
- `owt_valid.npy` — uint16 token IDs of 277 MB owt_valid.txt (66,458,625 tokens)
|
| 15 |
+
|
| 16 |
+
ID scheme matches `cs336_basics.tokenizer.Tokenizer` (special at 0).
|
| 17 |
+
|
| 18 |
+
## Quickstart on RunPod
|
| 19 |
+
```
|
| 20 |
+
pip install -U huggingface_hub
|
| 21 |
+
huggingface-cli download AndrewK101/cs336-owt-32k-bpe --repo-type dataset --local-dir data
|
| 22 |
+
```
|