The framework and its place.
WHAT IT IS
A framework for building, training and deploying machine learning models.
WHAT IT PROVIDES
Model construction, at high and low level Training, including across many machines Deployment to servers, browsers, mobile devices and embedded hardware
WHAT KERAS IS
The high-level interface for defining and training models, included with it.
WHAT TO USE
That interface, for almost everything.
WHY
The lower-level operations are rarely necessary.
WHAT THE ALTERNATIVE FRAMEWORKS ARE
Other frameworks are widely used, particularly in research.
WHAT DISTINGUISHES THIS ONE PRACTICALLY
Deployment breadth: server, browser, mobile and embedded from one ecosystem.
WHAT THAT MATTERS FOR
Getting a model into an application, which is where most projects fail.
WHAT TO ESTABLISH BEFORE STARTING ANY PROJECT
What decision the model informs How you would know it was working What data exists
WHAT MOST PROJECTS DISCOVER
The data is the problem, not the model.
WHAT TO EXPECT
That data preparation consumes the majority of the effort.
