ankit-pn/functiongemma-270m-phone-assistant
Full fine-tune vs google/functiongemma-270m-it. Fine-tune: 100% of the weights changed, across 18 of 18 layers.
Tensors changed: Embeddings 1/1, Norms 89/109, MLP 54/54, Attention 72/72
Full fine-tune vs google/functiongemma-270m-it. Fine-tune: 100% of the weights changed, across 18 of 18 layers.
Tensors changed: Embeddings 1/1, Norms 89/109, MLP 54/54, Attention 72/72