Let's now summarize all the key topics we've discussed in this course. Feel free to download the overview material in the end of this page.

Tensorflow Set Up

Instalation

Import

Tensor Types

Simple Tensor Creation

Tensor Properties

• Rank: It tells you the number of dimensions present in the tensor. For instance, a matrix has a rank of 2. You can get the rank of the tensor using the .ndim attribute:

• Shape: This describes how many values exist in each dimension. A 2x3 matrix has a shape of (2, 3). The length of the shape parameter matches the tensor's rank (its number of dimensions). You can get the the shape of the tensor by the .shape attribute:

• Types: Tensors come in various data types. While there are many, some common ones include float32, int32, and string. You can get the the data type of the tensor by the .dtype attribute:

Tensor Axes

Applications of Tensors

• Table Data

• Text Sequences

• Numerical Sequences

• Image Processing

• Video Processing

Batches

Tensor Creation Methods

Convertions

• NumPy to Tensor

• Pandas to Tensor

• Constant Tensor to a Variable Tensor

Data Types

Arithmetic

• Addition

• Subtraction

• Element-wise Multiplication

• Division

Broadcasting

Linear Algebra

• Matrix Multiplication

• Matrix Inversion

• Transpose

• Dot Product

Reshape

Slicing

Modifying with Slicing

Concatenating

Reduction Operations

Gradient Tape

@tf.function

What role does a loss function play in a neural network?

Selecione algumas respostas corretas