FWF, short for fine wine furry, a autismXL based model
Yet another furry model, trained on a top artist style on e621, with the their respectively more recent images for fine quality
Training info
Total steps: 1500
Total dataset images: 50
Training method: Diffusers
Convert tool: convert_diffusers_to_original_sdxl.py
Other training parameters
Learning rate: 4e-5
Batch size: 2
Gradient accumulation steps: 2
GPU used: H100 PCIE
Prompt guide
Main triggers: edjit, 2024/2025/2026, hi res, digital media (artwork)
Best resolution: 1280x960
For scoring: score_9, score_8_up, score_7_up, score_6_up
Negatives: deformed, incoherent, blurry, old, oldest
Examples
| Steps: 32 CFG:4 |
|---|
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support