Adapter 3: white cartoon portrait -> three-view character turnaround sheet (Qwen-Image-Edit-2511 LoRA)
Trained on maxbsoft/synth-v1-packed-a3 (131 synthetic pairs, 1344x768: portrait on a white canvas -> front / three-quarter / profile sheet).
No real people in the training data. Rank 32, lr 1e-4, 1000 steps. Chosen checkpoint: adapter3_portrait2sheet_qwen.safetensors (final, step 1000). See RESULT.md.
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support
Model tree for maxbsoft/cbai-adapter3-portrait2sheet-qwen2511
Base model
Qwen/Qwen-Image-Edit-2511