Transformers
deb-bpe / tokenizer_config.json
thedeba's picture
Add bilingual Bengali-English BPE tokenizer (100k vocab, byte_fallback)
9a0fd03 verified
Raw
History Blame Contribute Delete
216 Bytes
{
"backend": "tokenizers",
"bos_token": "[BOS]",
"eos_token": "[EOS]",
"is_local": true,
"model_max_length": 2048,
"pad_token": "[PAD]",
"tokenizer_class": "TokenizersBackend",
"unk_token": "[UNK]"
}