Rx_Codex_Tokenizer / tokenizer_config.json
rxmha125's picture
Upload folder using huggingface_hub
4b5410e verified
raw
history blame contribute delete
379 Bytes
{
"tokenizer_class": "PreTrainedTokenizerFast",
"bos_token": "[BOS]",
"eos_token": "<|endoftext|>",
"unk_token": "[UNK]",
"pad_token": "[PAD]",
"additional_special_tokens": [
"<|user|>",
"<|assistant|>",
"<|fim_prefix|>",
"<|fim_middle|>",
"<|fim_suffix|>"
],
"model_max_length": 32768,
"clean_up_tokenization_spaces": false
}