- In the Project space menu, select Collections.
- Select a collection.
- Select Models.
- Select Create Model > Create new model.
- Enter a Name and select the Model type.
- Configure the model as needed using the following fields:
- Select your Positive labels. At least one positive label is required.
- Select your Negative labels.
- Select a Pretrained model if desired.
- Enter a Description to describe the category you want to detect.
- Select a Train dataset to restrict the model to train on a subset of data.
- Select an Exclude train dataset to exclude a subset of data from model training.
- Select a Prediction dataset to restrict the model to predict on a subset of data.
- Select Create.
Models
Create a model
You can create a model to apply to your data for various purposes, including classification and information extraction. Once you create a model, you can train it using either using instructions or by providing examples.
To create a model, complete the following procedure.