santacoder / tokenizer_config.json
loubnabnl's picture
loubnabnl HF staff
Switch from PreTrainedTokenizerFast to GPT2TokenizerFast and add eos_token & bos_token (#15)
47ad9f0
{
"errors": "replace",
"tokenizer_class": "GPT2TokenizerFast",
"bos_token": "<|endoftext|>",
"eos_token": "<|endoftext|>",
"model_max_length": 2048
}