30b model

#3
by breawerawer - opened

Hi, I have been able to get the 125m and 6.7b model versions to run with nothing more than the sample code in the read me. However, the 30b model gives errors such as:

KeyError: 'decoder.layers.13.self_attn_layer_norm.weight'
KeyError: 'decoder.layers.31.self_attn.q_proj.bias
KeyError: 'decoder.layers.27.fc1.bias'

Is this somthing silly I am doing wrong or is this a bug?

Thanks so much!

Facing the same problem right now.

Same. Let's continue the discussion here:
https://huggingface.co/facebook/galactica-30b/discussions/4#637c8606d55081513c5679ef

The more hashsums of the blobs we have the better :)

Sign up or log in to comment