hyunseoki

/

ko-en-llama2-13b

like 26

Since 7b(https://huggingface.co/beomi/llama-2-ko-7b) model used 40B tokens for training, it would be helpful to let me know the total number of tokens for training 13b of yours.