sdxl-training
This is a LoRA derived from stabilityai/stable-diffusion-3-medium-diffusers.
The main validation prompt used during training was:
a naked woman, front view, standing in a room. large breasts, pretty face, pussy. photo, RAW candid cinema, 16mm, color graded portra 400 film, remarkable color, ultra realistic, dry skin, shot with cinematic camera
Negative prompt:
ugly 3d render, deformed corpse, brushstrokes, painting
Validation settings
- CFG:
4.0 - CFG Rescale:
0.0 - Steps:
40 - Sampler:
euler - Seed:
6 - Resolution:
816x1280
Note: The validation settings are not necessarily the same as the training settings.
You can find some example images in the following gallery:
- Prompt
- unconditional (blank prompt)
- Negative Prompt
- blurry, cropped, ugly
- Prompt
- a photo of a naked woman with large breasts
- Negative Prompt
- blurry, cropped, ugly
The text encoder was not trained. You may reuse the base model text encoder for inference.
Training settings
- Training epochs: 1072
- Training steps: 21450
- Learning rate: 0.0002
- Effective batch size: 20
- Micro-batch size: 5
- Gradient accumulation steps: 4
- Number of GPUs: 1
- Prediction type: epsilon
- Rescaled betas zero SNR: False
- Optimizer: AdamW, stochastic bf16
- Precision: Pure BF16
- Xformers: Enabled
- LoRA Rank: 64
- LoRA Alpha: 64.0
- LoRA Dropout: 0.1
- LoRA initialisation style: default
Datasets
curated3
- Repeats: 0
- Total number of images: 400
- Total number of aspect buckets: 1
- Resolution: 0.5 megapixels
- Cropped: False
- Crop style: None
- Crop aspect: None
- Downloads last month
- 15