snu-aidas/Dynin-Omni
Full fine-tune vs Gen-Verse/MMaDA-8B-MixCoT. Fine-tune: 100% of the weights changed, across 32 of 32 layers.
Tensors changed: Norms 65/65, Attention 128/128, other 64/64, MLP 32/32
Full fine-tune vs Gen-Verse/MMaDA-8B-MixCoT. Fine-tune: 100% of the weights changed, across 32 of 32 layers.
Tensors changed: Norms 65/65, Attention 128/128, other 64/64, MLP 32/32