yerevann/chemlactica-125m
Full fine-tune vs facebook/galactica-125m. Fine-tune: 100% of the weights changed, across 12 of 12 layers.
Tensors changed: Embeddings 1/1, Norms 50/50, MLP 48/48, Attention 96/96
Full fine-tune vs facebook/galactica-125m. Fine-tune: 100% of the weights changed, across 12 of 12 layers.
Tensors changed: Embeddings 1/1, Norms 50/50, MLP 48/48, Attention 96/96