shuttleai/shuttle-jaguar
Full fine-tune vs shuttleai/shuttle-3-diffusion. Fine-tune: 97% of the weights changed, across 0 of 0 layers.
Tensors changed: Norms 0/152, Attention 76/152, MLP 69/152, other 152/320
1 derivatives checked
- mit-han-lab/svdq-int4-shuttle-jaguar: Adapter