Update README.md
Browse files
README.md
CHANGED
|
@@ -1,3 +1,23 @@
|
|
| 1 |
---
|
| 2 |
license: mit
|
|
|
|
|
|
|
| 3 |
---
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
---
|
| 2 |
license: mit
|
| 3 |
+
language:
|
| 4 |
+
- en
|
| 5 |
---
|
| 6 |
+
|
| 7 |
+
## Baby Tokenizer
|
| 8 |
+
|
| 9 |
+
Compact sentencepiece tokenizer for sample-efficient English language modeling.
|
| 10 |
+
|
| 11 |
+
### Data
|
| 12 |
+
|
| 13 |
+
This tokeniser is derived from the BabyLM 100M dataset of mixed domain data, consisting of the following sources:
|
| 14 |
+
- CHILDES (child-directed speech)
|
| 15 |
+
- Subtitles (speech), BNC (speech)
|
| 16 |
+
- TED talks (speech)
|
| 17 |
+
- children's books (simple written language).
|
| 18 |
+
|
| 19 |
+
### Specifications
|
| 20 |
+
|
| 21 |
+
- Vocabulary size: 20k
|
| 22 |
+
- Alphabet limit: 150
|
| 23 |
+
- Minimum token frequency: 5
|