How did you resolve `llama_model_load: error loading model: missing tensor 'blk.92.attn_norm.weight'`?
#1
by
jivank - opened
Hi, thank you for this.
I saw, when searching for the error above, you had also run into it. Could you share how or if you were able to resolve it? I've tried pruning the layer, which simply made an error on layer 91. I have looked online but only have found suggestions to upgrade llama.cpp which I have already done. Please let me know, thanks again.
Hi, I have a discussion about it
https://huggingface.co/Ex0bit/GLM-4.7-PRISM/discussions/3
The solution was to change num_nextn_predict_layers from 1 to 0 in the config.json