metadata

license: creativeml-openrail-m
base_model: toilaluan/turbox
tags:
  - stable-diffusion
  - stable-diffusion-diffusers
  - text-to-image
  - diffusers
  - simpletuner
  - full
inference: true

full-training

This is a full rank finetune derived from toilaluan/turbox.

The main validation prompt used during training was:

ethnographic photography of teddy bear at a picnic

Validation settings

CFG: 7.5
CFG Rescale: 0.0
Steps: 30
Sampler: None
Seed: 42
Resolution: 1024

Note: The validation settings are not necessarily the same as the training settings.

The text encoder was not trained. You may reuse the base model text encoder for inference.

Training settings

Training epochs: 4
Training steps: 45
Learning rate: 8e-07
Effective batch size: 40
- Micro-batch size: 10
- Gradient accumulation steps: 4
- Number of GPUs: 1
Prediction type: epsilon
Rescaled betas zero SNR: False
Optimizer: AdamW, stochastic bf16
Precision: Pure BF16
Xformers: Not used

Datasets

xxx123

Repeats: 0
Total number of images: 360
Total number of aspect buckets: 1
Resolution: 1.0 megapixels
Cropped: False
Crop style: None
Crop aspect: None

Inference

import torch
from diffusers import DiffusionPipeline

model_id = 'full-training'
pipeline = DiffusionPipeline.from_pretrained(model_id)

prompt = "ethnographic photography of teddy bear at a picnic"
negative_prompt = "blurry, cropped, ugly"

pipeline.to('cuda' if torch.cuda.is_available() else 'mps' if torch.backends.mps.is_available() else 'cpu')
image = pipeline(
    prompt=prompt,
    negative_prompt='blurry, cropped, ugly',
    num_inference_steps=30,
    generator=torch.Generator(device='cuda' if torch.cuda.is_available() else 'mps' if torch.backends.mps.is_available() else 'cpu').manual_seed(1641421826),
    width=1152,
    height=768,
    guidance_scale=7.5,
    guidance_rescale=0.0,
).images[0]
image.save("output.png", format="PNG")