Sekai-1 / tokenizer_config.json
8BitStudio's picture
Add Sekai-1 chat model weights, tokenizer and How to use section
871f0c8 verified
Raw History Blame Contribute Delete
760 Bytes
{
"bos_token": "<|endoftext|>",
"clean_up_tokenization_spaces": false,
"eos_token": "<|im_end|>",
"model_max_length": 2048,
"pad_token": "<|endoftext|>",
"tokenizer_class": "PreTrainedTokenizerFast",
"unk_token": "<|endoftext|>",
"model_input_names": [
"input_ids",
"attention_mask"
],
"additional_special_tokens": [
"<|im_start|>",
"<|im_end|>",
"<|reserved_0|>",
"<|reserved_1|>",
"<|reserved_2|>",
"<|reserved_3|>",
"<|reserved_4|>"
],
"chat_template": "{{ bos_token }}{% for message in messages %}{{ '<|im_start|>' + message['role'] + '\n' + message['content'] + '<|im_end|>' + '\n' }}{% endfor %}{% if add_generation_prompt %}{{ '<|im_start|>assistant\n' }}{% endif %}"
}