This notebook covers some of the data preparation required, as well as training the model and evaluating the model.īecause this is a new notebook, you need to load the TensorFlow data again, as shown in Figure 16.įigure 16. Open the 02-MNIST-Tensorflow.ipynb notebook. Reloading the data and creating DataFrames for testing and training Creating a TensorFlow model with several convolution layers.Reloading the data and creating DataFrames for testing and training.The sections that you will be working through include: Some other approaches involve decision trees or support vector machines. AI also has the secondary benefit of being significantly easier to program in some cases. Although it is possible for non-AI code to do things such as classifying handwritten digits classification, AI is currently state of the art for such loosely defined tasks. In this learning path, we will use Keras to work on the MNIST data set. Instead, Keras requires just a general understanding of when to apply certain techniques. The library doesn't require a lot of the advanced math that some lower layers might need. Keras lets you look at neural networks in terms of layers of nodes and is generally easy for new users to use. It has several layers, allowing you to get as deep into the weeds as you need when writing code for machine learning.įor the purposes of this tutorial, we will stay at a fairly high level, using the packaged Keras library. TensorFlow, a machine learning library from Google, is the most well-known and widely used framework to do this kind of work. To really dive into AI, you need to use one of the many frameworks provided for these tasks. Build, train, and run your TensorFlow model
0 Comments
Leave a Reply. |