The newly launched Embarcadero Open-Source Project Stream aims to highlight some of the open-source projects that are sponsored by Embarcadero Technologies. The series kicked off with Tkinter which is a de-facto standard GUI for Python and also works well as windows ui toolkit as well as for other platforms. The second webinar was about PyTorch for Delphi and the new Python libraries such as the Python Environments, Lightweight Wrappers, and Python Data Sciences.
How does this PyTorch for Delphi Demo work?
This video below is an excerpt from the recently concluded webinar where we will get to see Python Data Sciences in action. What we have here is a PyTorch demo, a sample that is freely available on GitHub. Just be mindful that in order for the sample demo to work, you need to install all the necessary packages and Python prerequisites including the lightweight Python Wrappers, Python Environments, and Python4Delphi framework.
Once installed and once the server is up and running, the application (PyTorch Thumbs Up and Down demo) allows you to collect data or images using the phone’s camera. After collecting enough data (images), you will have the server run the training. Once the model is successfully trained, you can now proceed to the Live Recognition that can easily tell or identify the hand gestures taken from your camera whether you are making a thumbs up or a thumbs down.
Fascinating indeed! If you want to know more about how this PyTorch demo works, feel free to watch the video below. Take note that this is an early access preview, but you are encouraged to try it out and play around with these PyTorch demos.
Are you ready to try the PyTorch demo for yourself? Download your own personal copy of Delphi today!