pythia-1.4b-deduped-jokes / tokenizer_config.json
AlekseyKorshuk's picture
Training in progress, step 400
a235a3a
{
"add_prefix_space": false,
"bos_token": "<|endoftext|>",
"clean_up_tokenization_spaces": true,
"eos_token": "<|endoftext|>",
"model_max_length": 1000000000000000019884624838656,
"tokenizer_class": "GPTNeoXTokenizer",
"unk_token": "<|endoftext|>"
}