Chat Usage Error - "Input validation error"

#41
by alksndrmnts - opened

Hello!
I encountered a problem while using the chat, the following error occurs:
"Input validation error: inputs' tokens + max_new_tokens' must be <= 8192. Given: 7493 inputs' tokens and 2047 'max_new_tokens'."
I am trying to send a request which seems to exceed the allowed number of tokens. Can anyone suggest how I can solve or circumvent this issue? I am particularly interested in advice from those who have faced a similar error and managed to resolve it. Perhaps there are some techniques for reducing the request size or ways to optimize data input that would help avoid exceeding the limit.
Thank you for any help and advice!

Same problem here bur with small requests...

Sign up or log in to comment