AuraFlow v0.3 PVC style LoRA

Model description

PVC style LoRA for AuraFlow v0.3. Only tested with ComfyUI, and I don't know if this LoRA works on other WebUIs.

Trigger words

You should use pvc to trigger the image generation.

You should use pvc figure to trigger the image generation.

You should use figma to trigger the image generation.

You should use nendoroid to trigger the image generation.

Prompt guide

Long danbooru tags style prompt is preferred.

This LoRA was trained with danbooru tags, but not of all danbooru tags would work. Please consider to use natural language to align. (but too long natural language prompts sometimes break outputs.)

Training settings

Training tool: auraflow-ft
Base model: AuraFlow-v0.3-bnb-nf4
- Quantization: bitsandbytes NF4
  - Target keys: all except the exclude keys
  - Exclude keys: ["t_embedder", "final_linear", "modF"]
Peft method: LoRA
- LoRA settings
  - Rank: 16
  - Alpha: 1
  - Zero-init: lora_down
  - Target keys: [".attn.", ".mlp.", ".modC.", ".modC.", ".modX."]
  - Exclude keys: ["text_encoder", "vae", "t_embedder", "final_linear"]
Optimizer: RAdamScheduleFree from schedulefree
Learning rate: 1e-3
- Learning rate scheduler: not used
Loss type: L2
Timestep sampling: Logit-norm
Train epochs: 10
Dataset
- Num images: 7452
- Num repeats: 2
- Batch size: 2
- Bucket settings: (maybe buggy)
  - Maximum resolution: around 1024x1024
  - Minimum resolution: 384x384
  - Bucket size step: 128
- Caption augumentation
  - Shuffle tags: Yes
  - Tag dropout: No
  - Caption dropout: No
Gradient checkpointing: Yes
Training device: 1x RTX 4070Ti Super (16GB VRAM)
Training hours: around 62 hours
Checkpoint history: https://huggingface.co/p1atdev/afv03-lora/tree/main/pvc-2

Download model

Weights for this model are available in Safetensors format.