Supervised Contrastive Learning

Supervised Contrastive Learning

Lets you try a Keras model built for supervised contrastive learning in a Gradio page, in a browser or on your own computer. Aimed at students and developers who want to see it work.

Made by keras-io on Hugging Face. These notes are Mindket's, written from its Space page and README; the app's own files and license are on Hugging Face.

About

What it is: This Gradio demo from keras-io wraps one model, keras-io/supervised-contrastive-learning-cifar10. The model name points to an image model trained on CIFAR-10 with supervised contrastive learning. You try it in a browser, or you run a copy on your own computer.

Who it suits: It suits students and developers who want to see a Keras image model at work before they read any training code. Teachers can show it as a live demo.

Industry | Who uses it | How they use it | Example
Education | Deep learning students | Test a trained image model without writing code | Open the Space and try one of the example files
Research | Machine learning researchers | Look at how a finished model is wired into a demo | Read app.py next to the model name
Software | Python developers | Run a copy locally and adapt it | Download the Space and start app.py

Skip it if you need a training script or a general image tool, because this Space only wraps one model.

Where to use it:

• A web browser: open the Space page on Hugging Face and try it with nothing to install
• Windows, macOS or Linux: run a copy with Python

Use case | Who | What they get back
Try the model | Students | The model's answer on the page for the input they pick
Read the code | Developers | The app.py file that connects the model to a Gradio page
Teach with it | Instructors | A demo they can run on their own computer

How to set it up: You need a few things in place first.

• Python with pip, on Windows, macOS or Linux
• An internet connection to fetch the files from Hugging Face

1. Install the Hugging Face hub package.
2. Download the Space into a local folder.
3. Move into that folder.
4. Install Gradio and the packages listed in requirements.txt.
5. Start the app.

pip install -U huggingface_hub
hf download keras-io/supervised-contrastive-learning --repo-type space --local-dir supervised-contrastive-learning
cd supervised-contrastive-learning
pip install gradio
pip install -r requirements.txt
python app.py

How to use it: Online, open the Space page and use the Gradio interface it shows. On your own computer, run the main file from the project folder and open the page it serves in your browser.

python app.py

Example: You run python app.py inside the project folder. The Gradio page starts, you give it an input such as one of the files in the examples folder, and the page shows what the model returns for it.

For newcomers and experts: If this is new to you, start online. Open the Space page and try an example file before you install anything. It is the quickest way to see a Keras model working.

If you know the field, note that the Space names a single model and app.py loads it. Supervised contrastive learning uses class labels when it compares examples during training. This Space looks like inference only, going by its listed files, so it shows no training and gives no accuracy figures.

Open the Space or run app.py
The Gradio page loads
You give it an input
The named model scores it
The result appears on the page

Good to know: The source gives no accuracy numbers or training details, so test it on your own examples before you rely on it.

• License: apache-2
• Cost to run: none named, no API key, account or paid service
• Data it reaches: the files in the repository and the model it names, with nothing else listed

On Hugging Face, 13 November 2026

• Likes: 5
• Built with: gradio
• License: apache-2.0

Links

• Space: https://huggingface.co/spaces/keras-io/supervised-contrastive-learning
• Files: https://huggingface.co/spaces/keras-io/supervised-contrastive-learning/tree/main

Try it

Open the Space in your browser, or copy it to your own Hugging Face account from its page. Nothing in this kit runs by itself.