Upload README.md with huggingface_hub
Browse files
README.md
ADDED
|
@@ -0,0 +1,38 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
# Super Vocabulary
|
| 2 |
+
|
| 3 |
+
A merged super-vocabulary built from 21 tokenizer(s).
|
| 4 |
+
|
| 5 |
+
**Vocab size:** 122554
|
| 6 |
+
|
| 7 |
+
## Tokenizers
|
| 8 |
+
|
| 9 |
+
- `flexitok/bpe_arb_Arab_8000`
|
| 10 |
+
- `flexitok/bpe_ces_Latn_8000`
|
| 11 |
+
- `flexitok/bpe_ltr_cmn_Hani_8000_v2`
|
| 12 |
+
- `flexitok/bpe_dan_Latn_8000`
|
| 13 |
+
- `flexitok/bpe_deu_Latn_8000`
|
| 14 |
+
- `flexitok/bpe_ell_Grek_8000`
|
| 15 |
+
- `flexitok/bpe_fas_Arab_8000`
|
| 16 |
+
- `flexitok/bpe_fra_Latn_8000`
|
| 17 |
+
- `flexitok/bpe_fw_edu_8000`
|
| 18 |
+
- `flexitok/bpe_hun_Latn_8000`
|
| 19 |
+
- `flexitok/bpe_ind_Latn_8000`
|
| 20 |
+
- `flexitok/bpe_ita_Latn_8000`
|
| 21 |
+
- `flexitok/bpe_ltr_jpn_Jpan_8000_v2`
|
| 22 |
+
- `flexitok/bpe_nld_Latn_8000`
|
| 23 |
+
- `flexitok/bpe_pol_Latn_8000`
|
| 24 |
+
- `flexitok/bpe_por_Latn_8000`
|
| 25 |
+
- `flexitok/bpe_rus_Cyrl_8000`
|
| 26 |
+
- `flexitok/bpe_spa_Latn_8000`
|
| 27 |
+
- `flexitok/bpe_swe_Latn_8000`
|
| 28 |
+
- `flexitok/bpe_ltr_tur_Latn_8000_v2`
|
| 29 |
+
- `flexitok/bpe_vie_Latn_8000`
|
| 30 |
+
|
| 31 |
+
## Files
|
| 32 |
+
|
| 33 |
+
- `super_vocab.json` — merged vocabulary mapping token string → super index
|
| 34 |
+
- `config.yaml` — model config with `vocab_size`
|
| 35 |
+
- `participating_tokenizers.json` — list of tokenizer names included
|
| 36 |
+
- `<tokenizer>_super_mapping.json` — per-tokenizer index → super index mapping
|
| 37 |
+
- `<tokenizer>_vocab.json` — per-tokenizer vocabulary
|
| 38 |
+
- `<tokenizer>_info.json` / `.yaml` — tokenizer metadata
|