Upload README.md with huggingface_hub
Browse files
README.md
ADDED
|
@@ -0,0 +1,22 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
license: apache-2.0
|
| 3 |
+
tags:
|
| 4 |
+
- sentencepiece
|
| 5 |
+
- tokenizer
|
| 6 |
+
---
|
| 7 |
+
|
| 8 |
+
# North Tokenizer
|
| 9 |
+
|
| 10 |
+
Shared SentencePiece tokenizer for the North Star model family:
|
| 11 |
+
- [North Air 1](https://huggingface.co/arthu1/north-air-1) — 124M, fast
|
| 12 |
+
- [North Star 1](https://huggingface.co/arthu1/north-star-1) — 198M, balanced
|
| 13 |
+
- [Wind Arc 1.5](https://huggingface.co/arthu1/wind-arc-1.5) — 198M, deep reasoning
|
| 14 |
+
|
| 15 |
+
**Vocabulary size:** 32,000
|
| 16 |
+
**Format:** SentencePiece (`.model` file)
|
| 17 |
+
|
| 18 |
+
```python
|
| 19 |
+
import sentencepiece as spm
|
| 20 |
+
sp = spm.SentencePieceProcessor(model_file="tokenizer.model")
|
| 21 |
+
ids = sp.encode("Hello world", out_type=int)
|
| 22 |
+
```
|