课程: Building and Deploying Deep Learning Applications with TensorFlow

今天就学习课程吧！

今天就开通帐号，24,100 门业界名师课程任您挑！

Add custom visualizations to TensorBoard

Add custom visualizations to TensorBoard - TensorFlow教程

课程: Building and Deploying Deep Learning Applications with TensorFlow

开始一个月试用 购买团队方案

Add custom visualizations to TensorBoard

“

- [Instructor] TensorBoard allows you to create custom visualizations beyond just line graphs. You can use these visualizations to monitor your machine learning model and what kind of data it's generating. Currently TensorFlow supports these types of visualizations. First, the image visualization allows you to see any array of data as an image. You create an image visualization by adding a tf.summary.image object to your graph and passing it the array you want to visualize. This is helpful when you're building a neural network that classifies or generates images. In this example I was training a neural network to generate new images from a data set. Using the image visualization in TensorBoard, I was able to monitor the images the system was generating during the training process. You can also listen to audio data in TensorBoard. To add an audio player to TensorBoard, you create a new tf.summary.audio object and you add it to your computational graph, this is typically used when…

内容