Training Custom Models

Create custom models for image generation simply by uploading images

Webring.AI image generation supports training Custom Models, which allow users to generate images using a specific style or character based off of a set of training images.

Here are just a few examples of custom models in action:

Fine tuning a model with Webring.ai is easy. You just need a handful of images to get started. A custom model can be trained from as few as 25 images, although using more images produces more consistent models.

Step 0: Prerequisites

Here's what you'll need to get started:

• A Webring.ai Account

• At least 20-30 training images

• At least 150 credits (trainings cost 1 credit for every 2MB of training images).

Step 1: Collect your training images

You'll need a few images of your subject to get started. These should be high-quality images of the character or style you wish to train. The more variety among your images, the more flexible your model will be. The larger the # of images you use, the more consistent your model will be.

For best results you'll want to use at least 10 images or more. In theory you'll get continually better results as you include more images in the training data, but the training process will take longer (and cost more credits) the more images you add. Webring.ai currently enforces a maximum of 200 images / 500 MB per training data set.

Consider the following when gathering your training images:

• WebP, JPG, and PNG formats are all supported.

• Use 1024x1024 or higher resolution if possible.

• Filenames don't matter. Name your files whatever you like.

• Aspect ratio doesn't matter. Images can be square, landscape, portrait, etc.

• 20 images is a good minimum.

Variety is key. For best results, choose training images with different angles and content.

You can either zip this file (easiest) or upload the images one-by-one to Webring.ai.

Step 2: Choose a unique trigger word

Whenever you train an image model, you also choose a unique "trigger word" that you'll use later in your text prompts when generating images:

photo of YOUR_TRIGGER_WORD_HERE looking rad, riding on a chopper motorcycle

Here are some things to consider when choosing a trigger word:

• It should be something unique like MY_UNIQ_TRGGR. Think "vanity license plates", but without any length limits.

• It should not be an existing word in any language, like dog or cyberpunk.

• Case doesn't matter, but capital letters can help visually distinguish the trigger word from the rest of the text prompt.

Got your trigger word? Hold it in your mind for a second. You'll use it in the next step.

Step 3: Create and train a model

Navigate to https://webring.ai/user/models (log in if you need to).

Fill out the model form:

• Model Name (can be anything but must be unique)

• Model Description

• Trigger Word

Now upload your training images. Select Upload Zip in case you compressed the folder of images yourself. Otherwise you can choose Select Images and upload each image one by one.

Once your images are selected, click Create Model.

Once your model starts training you will see it pop up under Your Models at the bottom of the page.

The status will show as startingand then processingwhile the training proceeds.

You will see status messages appear under and periodically update the status as training proceeds.

Example Prompts will also populate for the model, generated automatically from the training images.

Training will take approx 10-15 minutes to complete, depending on the number of images in your training set and their file size.

Once training is complete you should see a succeededmessage. If you see anything else, feel free to reach out for help in our Discord server.

Step 4: Create an Image Generation Command

After your model is ready, you can create a new Image Generation Command for one of your agents by clicking the New Command button.

Select the agent for which to attach the image generation command. If you don't have an agent at this point you'll want to create one. Click Create New Image Generation Command after your agent is selected.

This will bring you to the Create Command page. Update the command name if you prefer. When you are done simply click "Create".

The Model will already be set to "Flux" and Custom Model will be set to the model you just created.

You will now see the Update Command page for the new command. Toggle Enable Playground to test it out.

Now try out a new prompt! Some tips:

• Take a look at the Example Prompts generated from your model on the models page to get inspiration for new prompts

• The more you use the trigger word for your model in the prompt, the more it will influence the AI to generate images influenced by your training set.

Step 5: Connect your Agent to Socials

You can now connect this command to your agent's socials, allowing Telegram / Discord / Twitter users to trigger the command via its name, as well as enabling the agent to generate content for Twitter posts.

You can always iterate and create new models using the best generations from your previous models. Have fun and share your generations with the community on Discord and Twitter.