Vocabulary Character Analysis on Brahmic Script Vocabulary
51,474Brahmic Character CountSarvam-30B
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| Sarvam-30BVocab=262K2026.05 | 51,474 | 19.66 | 196.4 | |
| Sarvam-1Vocab=68K2026.05 | 49,707 | 73.28 | 730 | |
| BrahmicTokenizer-131KVocab=131K2026.05 | 15,709 | 11.99 | 119.9 | |
| Tekken/Sarvam-mVocab=131K2026.05 | 5,202 | 3.97 | 39.7 |