November 2019
Intermediate to advanced
296 pages
7h 52m
English
The operation of calling TensorFlow.js is performed synchronously. This means that none of the operations can block the main thread of the application. This ensures that the interactive interface of the application can work smoothly. However, this does not indicate that the result is immediately ready because the actual computation is deferred by the backend of TensorFlow.js. The returned tensor is just a pointer to the tensor. The underlying data is prepared by the backend asynchronously.
Thanks to the asynchronous design of TensorFlow.js, programmers don't need to care about the asynchronous side of things as much. But to ensure that you don't block the main application thread and user interface interaction, ...
Read now
Unlock full access