“TensorFlow programs are usually structured into a construction phase, that assembles a graph, and an execution phase that uses a session to execute ops in the graph.”

What happen if we remove kernel_initialization in both Keras model and TensorFlow model? Do we really get the right answer? Are these results stable? What’s a potential cause to this? …

TENSORFLOW Tensors: n-dimensional arrays Vector: 1-D tensor Matrix: 2-D tensor Flow: data flow computation framework A sequence of tensor operations

Building a Keras TensorFlow-Based Model 146

TensorFlow comes with its own implementation of this API, called tf.keras, which provides support for some advanced TensorFlow features (e.g., to efficiently load data).

In this paper, we describe the TensorFlow dataflow model and demonstrate the compelling performance that Tensor-Flow achieves for several real-world applications.

a mental map of the model’s architecture. Since diagrams are critical for their work, machine learning experts desire a tool that can auto-matically visualize the model structures.