Have you ever wanted to create your own machine learning models without writing any code? If so, you might be interested in Lobe, a free software that lets you build, train and test custom models using a simple drag-and-drop interface. In this blog post, I will run through the basics of Lobe and how you can use this model to make money by selling it as a service or an app.
What Is Lobe?
Lobe is a software that runs on your computer and uses your webcam or microphone to capture data for your model. You can also import data from files or online sources. Lobe supports image, sound and text data, and has built-in tools to help you label, filter and augment your data. You can also use Lobe to split your data into training, validation and test sets.
Once you have your data ready, you can use Lobe to design your model using a visual editor. You can choose from different types of layers, such as convolutional, recurrent or attention, and adjust their parameters, such as filters, kernels or activation functions. You can also add custom code blocks if you need more flexibility. Lobe will automatically generate the code for your model in Python or TensorFlow.
After you have designed your model, you can use Lobe to train it using your computer’s GPU or CPU. You can monitor the training progress and see the accuracy, loss and confusion matrix of your model. You can also use Lobe to test your model on new data and see how it performs.
When you are done, you can export your model as a TensorFlow Lite file, which can be used on mobile devices or embedded systems, or as an ONNX file, which can be used on various platforms and frameworks.
Let Jake from Lobe give you a detailed rundown of this great software in the video below.
If you want to try Lobe for yourself, you can download it for free from lobe.ai.
You can also find more tutorials and examples on their website.
