JuanCarlosMartinezSevilla/jazzmus-model
Full fine-tune vs antoniorv6/smt-grandstaff. Fine-tune: 100% of the weights changed, across 9 of 9 layers.
Tensors changed: Attention 128/128, MLP 32/32, Norms 78/80, Embeddings 5/5, other 113/113
Full fine-tune vs antoniorv6/smt-grandstaff. Fine-tune: 100% of the weights changed, across 9 of 9 layers.
Tensors changed: Attention 128/128, MLP 32/32, Norms 78/80, Embeddings 5/5, other 113/113