SurjoLabs/Surjo-50m-SFT-Only
Full fine-tune vs SurjoLabs/Surjo-50m. Fine-tune: 99% of the weights changed, across 10 of 10 layers.
Tensors changed: Embeddings 1/1, Norms 27/35, MLP 30/30, Attention 109/112
Full fine-tune vs SurjoLabs/Surjo-50m. Fine-tune: 99% of the weights changed, across 10 of 10 layers.
Tensors changed: Embeddings 1/1, Norms 27/35, MLP 30/30, Attention 109/112