Context length

#4
by ryanse - opened

Hi,

Is there a reason the tokenizer has a model_max_length of 16384?

from transformers import AutoTokenizer

tokenizer = AutoTokenizer.from_pretrained('deepseek-ai/DeepSeek-Coder-V2-Instruct-0724')
tokenizer.model_max_length
> 16384

Thanks!

Sign up or log in to comment