lm_training / tokenizer_config.json
jumelet's picture
distilgpt2
18b6c87
raw
history blame contribute delete
239 Bytes
{
"bos_token": "<s>",
"clean_up_tokenization_spaces": true,
"eos_token": "</s>",
"model_max_length": 1000000000000000019884624838656,
"pad_token": "<pad>",
"tokenizer_class": "PreTrainedTokenizerFast",
"unk_token": "<unk>"
}