Qwen2.5 icon indicating copy to clipboard operation
Qwen2.5 copied to clipboard

相同的程序,用 Qwen1.5-7B-Chat-GPTQ-Int4 没问题,用Int8则推理时报错

Open davidjia1972 opened this issue 11 months ago • 4 comments

相同的程序,用 Qwen1.5-7B-Chat-GPTQ-Int4 没问题,用 Qwen1.5-7B-Chat-GPTQ-Int8 在推理的时候报错:

RuntimeError: probability tensor contains either inf, nan or element < 0

davidjia1972 avatar Mar 23 '24 16:03 davidjia1972

this is the problem of autogptq. gotta report this to them

JustinLin610 avatar Mar 25 '24 02:03 JustinLin610

same here

i7990X avatar Mar 25 '24 07:03 i7990X

same to me

shikimoon avatar Apr 15 '24 03:04 shikimoon

same

Rundong-Li avatar Apr 22 '24 08:04 Rundong-Li