mytranslatenisa/ramt-labse-3
Full fine-tune vs mytranslatenisa/m2m100_combine_FineTuned. Fine-tune: 100% of the weights changed, across 12 of 12 layers.
Tensors changed: Norms 124/124, Attention 288/288, MLP 96/96, other 1/1
Full fine-tune vs mytranslatenisa/m2m100_combine_FineTuned. Fine-tune: 100% of the weights changed, across 12 of 12 layers.
Tensors changed: Norms 124/124, Attention 288/288, MLP 96/96, other 1/1