pycodegpt_trainer / tokenizer_config.json
hsultanbey's picture
hsultanbey/pycodegpt_trainer
bf906ee
raw
history blame contribute delete
291 Bytes
{
"add_prefix_space": false,
"bos_token": "<|beginoftext|>",
"clean_up_tokenization_spaces": true,
"eos_token": "<|endoftext|>",
"errors": "replace",
"model_max_length": 2048,
"pad_token": "<|padoftext|>",
"tokenizer_class": "GPT2Tokenizer",
"unk_token": "<|unkoftext|>"
}