Deep learning is one of those often mentioned yet not quite tangible topics for most people. If you’re unfamiliar with what deep learning is, check out this blog post or this deep learning guide. Don’t worry, we’ll wait for you to return.
This post will now assume you know about deep learning and will show you the deep capabilities for deep learning with Dataiku. See what I did there? But seriously, there are so many great opportunities to start your deep learning journey in Dataiku DSS as the platform that truly puts the power of AI in your hands. We’ll cover every aspect, from labeling your data to building your first model with visual machine learning (ML) or code notebooks to jumpstarting your process with transfer learning. TL;DR - If you’d rather just jump right into deep learning in Dataiku DSS check out these walkthroughs to build your first model, go deeper into natural language processing (NLP) or image classification.
Visual ML With a Side of Code
Dataiku provides a visual ML tool but will also require a little bit of coding skill to define the deep learning architecture using the Keras and TensorFlow libraries. You write the code that defines the architecture of your deep learning model and Dataiku DSS then handles the rest! From preprocessing your data to handling those missing values to training, deploying, and scoring the model, this becomes like any other model created and managed in Dataiku. Since deep learning models require so much compute power, Dataiku supports training on CPUs up to multiple GPUs and through container deployment capabilities. You can easily train and deploy models on cloud-enabled dynamic GPUs clusters.
With great power comes great responsibility so be sure to check out this step by step walkthrough to build your first deep learning model within Dataiku. There are also other walkthroughs leveraging the visual ML tools to help you on your natural language processing or image classification journey. Your first deep learning model is minutes away from being deployed!
Full-On Code Master
Next up are the unlimited capabilities of coding within Dataiku DSS. Get ready to code your heart out to build that custom deep learning model you’ve always dreamed of deploying because Dataiku does not restrict you to the algorithms that are part of its visual ML capabilities. You can code your own model, either in Python or in Scala, or start with the above visual ML and customize to build out your model for your unique needs. From Keras, TensorFlow, Tensorboard and more, the world is your oyster!
Use our code walkthroughs for natural language processing and image classification to see with support for the latest and greatest deep learning algorithms, libraries, and environments. Get to coding!
ML-Assisted Labeling
But wait, there’s more! If you are working with your own data, you’ll need a LOT of labeled data to tell a computer what the data is, which can be a big undertaking. If you don’t have labeled data yet, don’t worry! Dataiku makes this part easy with the free ML-assisted labeling plugin that helps you easily label your tabular, sound, or image datasets. For example, the image below shows object detection in action so you can clearly define what every object is in a picture.
Audio, Images & Videos, Oh My!
For nontraditional data like audio, you may have data but can’t figure out quite how to leverage it in deep learning. For any .wav files, the Speech to Text plugin will automatically transcribe your file to provide text that you can use to analyze in deep learning, sentiment analysis, etc.
For images and videos, you can leverage the Object Detection plugin which will detect the location and class of several objects in an image or video. For both images and videos, the plugin will automatically place bounding boxes around the objects and will produce a copy with it drawn on top of the video.
Transfer Learning
If you want to jumpstart your deep learning efforts with transfer learning, the Dataiku deep learning image plugin provides pre-trained models leveraging Keras and TensorFlow to score images and obtain classes, or for feature extraction. The models that are provided can be retrained to specialize on a particular set of images to kickstart your deep learning in Dataiku DSS journey.
There you have it — a basic overview of all of the ways you can achieve your deep learning objectives easily in Dataiku. Deep learning will continue to change the way we interact with the world around us, will you be part of this ML revolution?