codelion/dhara-250m
Full fine-tune vs codelion/dhara-250m-ar-base. Fine-tune: 100% of the weights changed, across 32 of 32 layers.
Tensors changed: other 64/64, Norms 0/129, MLP 128/128, Attention 224/256
Full fine-tune vs codelion/dhara-250m-ar-base. Fine-tune: 100% of the weights changed, across 32 of 32 layers.
Tensors changed: other 64/64, Norms 0/129, MLP 128/128, Attention 224/256