When we talk about backend implementation, we mainly talk about the topic of computation. But there is another interesting topic the surrounds data transfer and buffer management. Data transfer provides a form of input and fetches the result from computational units such as GPUs. How can we access the underlying buffer of the tensor? When can we get the result? Tensor management answers these questions. By definition, tensors are the fundamental building blocks of TensorFlow.js. Mastering how to use tensors efficiently allows us to write performant machine learning applications with the platform.
Tensor management
Tensor construction
A tensor is a data structure that holds metadata such as shapes, data types, and pointers...
