Skip to main content

The Flora App

Flora is an multimodal learning app built with PlayTorch. It was built through a collaboration between the PlayTorch team and PhD students in Stanford University's Human-Computer Interaction Graduate Program.

The app presents a sample lesson about eucalyptus trees in California and makes use of "multimodal learning" -- learning that happens through a variety of different modes like visual, kinaesthetic, reading, writing and speaking. PlayTorch enabled us to quickly integrate a variety of machine learning models that make the learning experience interactive and engaging.

The Flora App

Try it out​

You can try out Flora for yourself!

To get started follow these steps:

  1. Download the PlayTorch app onto an iOS or Android device. [TODO: Add link]
  2. Open the PlayTorch app and use it to scan the QR code on the right side of this Expo Snack [TODO: Add link].
  3. Start the lesson! Your first task is to find a eucalyptus tree and scan it using the custom-tuned imagenet classifier. If you don't have a eucalyptus tree nearby try using one of these pictures.

PlayTorch allowed our team to quickly integrate and use three different machine learning models, which you'll see as you engage with the app:

  1. The imagenet classifier for identifying tree types
  2. The wav2vec model, which enables high-quality speech-to-text.
  3. The BERT QA model for answering basic questions about eucalyptus trees.