How to avoid long prompt reasonable

Open techzhou opened this issue 3 months ago • 1 comments

deal with openai.InternalServerError: Error code: 500 - {'detail': '[address=0.0.0.0:43769, pid=1065] The decoder prompt (length 66501) is longer than the maximum model length of 65536. Make sure that max_model_len is no smaller than the number of text tokens.'}

back in the loop at least

Nov 23 '25 15:11 techzhou

Could you provide a fuller stack trace + invocation + context about what you're doing?

Nov 24 '25 14:11 eyurtsev