sea-sdxl-full

sea-sdxl-full

This is a full rank finetune derived from stabilityai/stable-diffusion-xl-base-1.0.

The main validation prompt used during training was:

A high‑resolution photo‑realistic image of a wild tiger in jungle

Validation settings

• CFG: 2.0
• CFG Rescale: 0.0
• Steps: 30
• Sampler: ddim
• Seed: 42
• Resolution: 1024x1024

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

<Gallery />

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

Training settings

• Training epochs: 14
• Training steps: 99
• Learning rate: 5e-06
  • Learning rate schedule: constant
  • Warmup steps: 100
• Max grad value: 1.0
• Effective batch size: 64
  • Micro-batch size: 16
  • Gradient accumulation steps: 2
  • Number of GPUs: 2
• Gradient checkpointing: True
• Prediction type: epsilon (extra parameters=['training_scheduler_timestep_spacing=trailing', 'inference_scheduler_timestep_spacing=trailing'])
• Optimizer: adamw_bf16
• Trainable parameter precision: Pure BF16
• Base model precision: no_change
• Caption dropout probability: 0.1%

Datasets

dragon-test-image

• Repeats: 0
• Total number of images: ~220
• Total number of aspect buckets: 1
• Resolution: 1.0 megapixels
• Cropped: True
• Crop style: random
• Crop aspect: square
• Used for regularisation data: No

seavl_crawling-test-image

• Repeats: 0
• Total number of images: ~200
• Total number of aspect buckets: 1
• Resolution: 1.0 megapixels
• Cropped: True
• Crop style: random
• Crop aspect: square
• Used for regularisation data: No

Inference

import torch
from diffusers import DiffusionPipeline

model_id = 'samuelcahyawijaya/sea-sdxl-full'
pipeline = DiffusionPipeline.from_pretrained(model_id, torch_dtype=torch.bfloat16) # loading directly in bf16

prompt = "A high‑resolution photo‑realistic image of a wild tiger in jungle"
negative_prompt = 'blurry, cropped, ugly'

pipeline.to('cuda' if torch.cuda.is_available() else 'mps' if torch.backends.mps.is_available() else 'cpu') # the pipeline is already in its target precision level
model_output = pipeline(
    prompt=prompt,
    negative_prompt=negative_prompt,
    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(42),
    width=1024,
    height=1024,
    guidance_scale=2.0,
    guidance_rescale=0.0,
).images[0]

model_output.save("output.png", format="PNG")

Exponential Moving Average (EMA)

SimpleTuner generates a safetensors variant of the EMA weights and a pt file.

The safetensors file is intended to be used for inference, and the pt file is for continuing finetuning.

The EMA model may provide a more well-rounded result, but typically will feel undertrained compared to the full model as it is a running decayed average of the model weights.