User Guide: Image Model Training & Generation App
Table of Contents
1. Usage Limits & Account Information
How Much Can I Create?
- 2 Apps: You can train up to two image models/apps.
- 50 Images: You can generate up to 50 images in total.
- 3 Videos: You can create up to three videos based on generated or uploaded images.
Where to Check Remaining Usage
Your remaining usage is displayed in the top-right corner of the screen, to the left of your account icon (the button that allows you to manage your account). This area provides a quick overview of how many apps, images, and videos you have left to create.
2. Creating Your Own Image Model
Step 1: Navigate to the "Create a New Image App" Section
Look for the "Create a New Image App" section at the top of the left column. This is where you will name your app and upload images to train your custom model.
Step 2: Name Your Image App
Enter a name for your app in the "Your App's Name" input field. Be creative! Examples include:
- "PetPortraitPainter"
- "StarryNightGenerator"
- "MyFamilyPortraitStyle"
There are no strict rules for naming. Choose something fun and memorable.
Step 3: Upload Images to Train the Model
You can upload images using one of the following methods:
- Upload a ZIP File: Click the "Upload ZIP" button. Select a `.zip` file containing multiple images. Maximum upload size is 10 MB.
- Upload Individual Images: Click the "Upload Images" button to choose one or more images directly from your device.
Step 4: Preview Your Uploaded Images
Once uploaded, your images will appear in the preview area below the upload options. Double-check that your uploads match your vision for the model.
Step 5: Start the Model Training
After uploading your images, click "Start Creating App". The training process will begin, and a progress indicator will appear. Once finished, a notification will confirm that your new app/model is ready to use.
Importing Models Made By Other Users
To use an image app made by other users you need to copy its address.
Step 1: Copy existing app’s address.
-
a) If you want to use the image app made by other users on this website and shared
in apps gallery, you can copy its address
(it would look like e.g.
galadriel:42d770944cf35d8da832c1ef80f8bd84cbc557548f8d2f98de831a652d134d16) via the copy icon on each app’s card. -
b) If you would like to import a third-party app, please copy its full HTTPS address that ends with
format=SafeTensor. For example, the link to the “Vividizer” app by Krawuzzn @civit.ai would be: https://civitai.com/api/download/models/742813?type=Model&format=SafeTensor
Step 2: Navigate to the "Import New App" Section in the right column.
Step 3: Enter the App’s Details
- Model Name: Choose any name you want for the imported app.
- Model Address: Paste the app’s address in the "App’s Address" field.
Step 4: Click “Import Model”
Step 5: Navigate to “Generate Image” section
Step 6: In the dropdown “Choose an Existing Model:” menu, your imported app should be highlighted.
Step 7: Choose the generation parameters (e.g., describe your desired image (prompt), choose aspect ratio, optionally set a “seed” for consistency).
Step 8: Click the “Generate Image” button.
Step 9: Your image will appear within a few seconds (but could take up to a minute) in the right column.
4. Generating Images
Step 1: Navigate to the "Generate Image" Section
Select a model to use for generating images. Choose your trained model, imported model, or the Base Model (Flux1.1-Pro).
Step 2: Configure Image Settings
Enter a prompt, select aspect ratio, and provide an optional seed.
Step 3: Generate the Image
Click "Generate Image". A loading animation will appear while your image is being generated.
5. Creating Videos from Images
Upload or select an existing image, provide a video prompt, and generate a video.
6. Tips for Best Results
- Naming Models: Be descriptive and creative with your names.
- Training Models: Use high-quality, consistent images for training.
- Generating Images: Write detailed prompts for specific results.