load error

by kl1996 - opened Oct 10, 2023

kl1996

Oct 10, 2023

i have a error of "ValueError: Trying to set a tensor of shape torch.Size([1792, 4096]) in "qweight" (which has shape torch.Size([3584, 4096])), this look incorrect." i don't know how to solve it

kl1996

Oct 10, 2023

@TheBloke

TheBloke

Owner Oct 10, 2023

Try downloading again - could be a bad download

JPBriem

Oct 26, 2023

@kl1996 , I have the same issue and downloading the model again doesn't help. Did you solve the problem?

For me, it only appeared for the "gptq-8bit-32g-actorder_True" version.

TeamResearch

Nov 27, 2023

Same issue as @JPBriem , only for the 8-bit model too

iarbel

Nov 29, 2023

Same for me. Happens in 32g & 128g

davidxmle

Apr 13

@TheBloke I am still getting the same issue for the two 8 bit versions... Was it during the remake something went wrong?

Borko24

Apr 23

I have the same error. Did anyone succeed in resolving it?

Upload images, audio, and videos by dragging in the text input, pasting, or clicking here.

Tap or paste here to upload images

· Sign up or log in to comment