1. Scenario Knowledge Center
  2. FAQs
  3. Choose an FAQ Based on Your Knowledge Level

Frequently Asked Questions (FAQ) After Training Your First Custom Model

Delve deeper into the nuances of model fine-tuning on Scenario.

If you haven't made your first Model yet, start here: Learning about Custom Models

Understanding Models

  - What is a Model

A tool that allows for the creation of specific styles or objects in images.

  - Training Page Management

You can close any active training or image generation page without halting progress.

  - Background in Training Images

Depending on the training's objective, removing the background might enhance outputs.

  - Image Dimensions for Training

All images should be squared and will be adjusted to 512x512 if training on Stable Diffusion 1.5 or 1024x1024 if training on SDXL by Scenario. Images outside these dimensions need adjustments. For more information see: Formatting Images for use on Scenario.

  - Improving Model Performance

Ensure a diverse set of unique images in your dataset, ranging from 5 to 50 images. The quality and diversity of images are pivotal.

  - Modifying a Model

If outputs aren't as expected, adjust your training dataset by copying the existing model and modifying the images.

  - Downloading CKPT Files

Currently, CKPT files aren't available for download, but this feature is on the horizon.