It is always interesting to see tools getting open sourced by giants like Facebook or Zalando.
Uber is one of those giants that capture our attention whenever they open source a tool and if there is something we absolutely love here at JAXenter, that is open source!
Today, we take a look at Ludwig, a toolbox that makes deep learning easier to understand for non-experts and promises to enable faster model improvement iteration cycles for experienced machine learning developers as well as researchers.
Let’s have a look at what makes this tool so interesting.
It’s all about writing zero code
As an analogy, if deep learning libraries provide the building blocks to make your building, Ludwig provides the buildings to make your city, and you can chose among the available buildings or add your own building to the set of available ones.
According to the official press release, Ludwig was originally designed as a generic tool for simplifying the model development and comparison process when dealing with new applied machine learning problems. The inspiration behind it comes from other machine learning software like Weka and MLlib, Caffe, and scikit-learn.
This unique mix brought to life this deep learning tool that aims to provide a set of model architectures that can be combined together to create an easy to use, end-to-end model.
Ludwig’s characteristics can be summarized as follows:
Code-free – No coding skills are required to train a model and use it for obtaining predictions.
Generality – A new data type-based approach to deep learning model design that makes the tool usable across many different use cases.
Flexibility – Experienced users have extensive control over model building and training, while newcomers will find it easy to use.
Extensibility – Easy to add new model architecture and new feature data types.
Understandability – Deep learning model internals are often considered black boxes, but we provide standard visualizations to understand their performance and compare their predictions.
How does it work, you ask? Well, Ludwig brings a new concept in the deep learning ecosystem – encoders map the raw data to tensors, and decoders map tensors to the raw data. This allows users to access combiners (glue components of the architecture) that combine the tensors from all input encoders, process them and return the tensors to be used for the output decoders.
Several input and output features may be specified in Ludwig’s model description file, and their combination covers many machine learning tasks. Source: Uber Engineering
Head over to the official website to have a look at some in-depth examples.
SEE ALSO: CosmoFlow lets astronomers harness the power of deep learning to learn the mysteries of deep space
Getting started
If you are looking to get started with Ludwig, keep in mind that there are several requirements you need to fulfill, so make sure to check out the official installation guide.
The post Uber’s Ludwig makes deep learning more understandable for amateurs and faster for experts appeared first on JAXenter.
Source : JAXenter
