Using Saagie Hugging Face Model Server as an App

You can use the Saagie Hugging Face Model Server’s user-friendly interface to quickly and easily deploy and predict Hugging Face deep learning models for text classification.

Before you begin:

Make sure the official Saagie repository is up to date. For more information, see Updating Repositories.
Verify that you have enabled the Saagie HF ModelServer TextCLF app in your project settings. If not, modify them. The app is available in the official Saagie repository.

Create the Saagie HF ModelServer TextCLF app by clicking Install an app from your project’s Apps page.
Your app appears in the project’s app library and is ready to go!
Click Open on your app card.
The Saagie HF ModelServer TextCLF app opens in a new tab.
In the Model Name field, enter the repository name of the Hugging Face model you want to deploy.

You can select a model directly from the Hugging Face Text Classification list. For example, the emotion classification model j-hartmann/emotion-english-distilroberta-base.
In the Label field, specify the output labels of the model.
Click Deploy and wait for loading to finish.
Once finish, your model is ready to use.

To deploy another model, update the Model Name and Label fields.
Enter your text in the Text Classification field to predict it. Each line break indicates the beginning of a new sentence, and thus a new prediction.
Click Predict to get the prediction results.
Your text prediction appears in the Inference Results (8) field.