TensorFlow is arguably the most popular machine learning (ML) framework today because of its rich multi-layer API. However, as a framework for ML modeling via code, TensorFlow can be a handful for beginners. Even experienced data scientists and developers can find it difficult when working with large sets of code to visualize the model, to see how changes to logic and hyperparameters affect the model, and to track down bugs.

Just released PerceptiLabs 0.11, is quickly becoming the GUI and visual API for TensorFlow that aims to solve these challenges. It’s built around a sophisticated visual ML modeling editor in which you drag and drop components and connect them together to form your model. PerceptiLabs automatically creates the underlying TensorFlow code, effectively wrapping that code inside of visual components so that you can easily visualize your model.

PerceptiLabs is a highly interactive tool. As components are added and configured, each displays a live preview to show how it processes input. There are also debugging tools, giving you tips, warnings, and errors while you are building your model so you can see where something is wrong and fix it right away.

PerceptiLabs gives you a choice of how you want to work. You can tweak hyperparameters in the visual modeling editor or you can modify the underlying code in PerceptiLabs’ code editor.

Categories of components include data sources, data processing (e.g., to reshape data), deep learning operations (e.g., convolution layers), math operations, and different training methods (e.g., classification, reinforcement learning, and many more).

#api #deep learning #perceptilabs #tensorflow