Semi-Supervised Contrastive Learning with SimCLR

Semi-Supervised Contrastive Learning with SimCLR

Runs an image classifier trained with SimCLR contrastive learning and few labels, from the Keras team, in a Gradio page with example images. For machine learning learners and researchers.

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 app runs an image classifier built with semi-supervised contrastive learning, using the SimCLR method. The record names one model, semi-supervised-classification-simclr, from the keras-io group. An examples folder holds sample images.

Who it suits: It suits machine learning learners and researchers who want to try a model trained with only some labeled data, and developers who want a small Gradio demo around a Keras model. The record does not list the classes or the dataset.

Industry | Who uses it | How they use it | Example
Research | Vision researchers | Try a model trained with few labels | Run an example image and read the prediction
Education | Machine learning students | See what semi-supervised training can produce | Compare predictions across several examples
Software | Python developers | Read a small Gradio demo around a Keras model | Open app.py and requirements.txt

You can skip it if you need a model with documented classes and accuracy.

Where to use it:

• A web browser, on the Space's page on Hugging Face
• Windows, macOS or Linux with Python, from a downloaded copy

Use case | Who | What they get back
Classify an example image | Learners | The model's predicted label
Try the bundled examples | Anyone | A quick sense of how the model behaves
Reuse as a base | Developers | A small Gradio app around a Keras model

How to set it up:

• Nothing at all if you use the online Space
• Python on your computer, only if you run a copy
• The packages listed in requirements.txt

1. Open the Space's page in a browser to use it with nothing to install.
2. To run a copy, install the Hugging Face Hub package and download the Space.
3. Move into the new folder, then install Gradio and the listed requirements.
4. Start the app with Python.

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

How to use it: Online, open the Space and use the controls on its page. On your own machine, run the main file from the Space folder, then use the page it serves. The example images are the quickest way to try it.

python app.py

Example: You open the Space and pick one of the example images. The app returns the label the model predicts for it. The record does not describe the output, so look at the page.

For newcomers and experts: Labeling images by hand is costly. Semi-supervised training uses a small set of labeled images together with many unlabeled ones. Start by opening the Space and running an example image.

If you know the field, SimCLR trains an encoder by pulling together two altered views of the same image and pushing apart views of different images. A classifier is then fitted on few labels. The record gives no dataset, label counts or accuracy figures.

You pick or upload an image
The Gradio page sends it to app.py
The trained encoder and classifier process it
The page shows the predicted label

Good to know:

• License: Apache 2, as listed on the Space
• Cost to run: none named
• Data it reaches: the images you give it and the model it names
• Not for: the record names no exclusions, and it gives no dataset or accuracy figures

On Hugging Face, 13 November 2026

• Likes: 1
• Built with: gradio
• License: apache-2.0

Links

• Space: https://huggingface.co/spaces/keras-io/semi-supervised-classification
• Files: https://huggingface.co/spaces/keras-io/semi-supervised-classification/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.