You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
It's possible that llama3 doesn't stop generation in time due to the tokenization change. Though we have updated several examples (e.g., CoT), we haven't got time to go through every example in our library. Could you try looking at the tokenization if it's urgent?
It's possible that llama3 doesn't stop generation in time due to the tokenization change. Though we have updated several examples (e.g., CoT), we haven't got time to go through every example in our library. Could you try looking at the tokenization if it's urgent?
yes, the problem fixed when i replace the eos_id. thanks
llm: llama3
rap strageqa
the first question takes more than 1 hours runing on A100
The text was updated successfully, but these errors were encountered: