mentee-embed-v3 / tokenizer_config.json
SyedSyab's picture
Fix tokenizer special tokens: use [PAD]/[UNK] from existing vocab
6f035c1 verified
Raw
History Blame Contribute Delete
293 Bytes
{
"tokenizer_class": "MenteeTokenizer",
"auto_map": {
"AutoTokenizer": [
"tokenization_mentee.MenteeTokenizer",
null
]
},
"model_max_length": 512,
"padding_side": "right",
"truncation_side": "right",
"pad_token": "[PAD]",
"unk_token": "[UNK]"
}