FFusion/FFXL400

Full fine-tune vs FFusion/FFusionXL-BASE. Fine-tune: 100% of the weights changed, across 12 of 12 layers.

Tensors changed: Embeddings 591/591, other 658/662, Attention 716/716, Norms 546/546

2 derivatives checked

• FFusion/400GB-LoraXL: Adapter
• aegon-h/SDXL-LORA: Adapter

On the Hub · Downloads on Model Pulse