Shulou(Shulou.com)06/01 Report--

This article mainly introduces the relevant knowledge of how to install TensorWatch, the content is detailed and easy to understand, the operation is simple and fast, and has a certain reference value. I believe you will gain something after reading this article on how to install TensorWatch. Let's take a look at it.

Microsoft has released an artificial intelligence debugging and visualization tool called TensorWatch on the GitHub open source platform. TensorWatch can be used to create custom visualization, UI, and dashboards. It can also be used to execute any query against your real-time ML training process, return a stream as the result of the query, and view the stream using the visualization tool of your choice.

TensorWatch is a Python library that uses Jupyter Notebook instead of a prepackaged user interface, based on which Jupyter Notebook is easier to customize. When conducting interactive debugging of the real-time training process, you can choose to use the composable UI in Jupyter Notebook or the real-time shared dashboard in Jupyter Lab. You can also use Python to build your own custom UI, or select an existing Python UI. TensorWatch also supports a variety of standard visualization types, including bar, bar, and pie charts, and supports 2D and 3D versions.

Data in TensorWatch is treated as a stream, and other objects include files, consoles, sockets, cloud storage, and visualization. TensorWatch streams can listen to other streams so that custom data flow graphs can be combined. Developers say this means that you can render many streams as the same visualization, or a stream can be rendered as many visualizations at the same time, or a stream can be persisted into many files, or not persisted at all.

Another major idea used in TensorWatch is the deferred logging mode. This means that you don't have to set up explicit logging; instead, you can have TensorWatch observe the variables. This reduces overhead, so you can track any number of variables as needed, including large models or entire batches of products during training.

You can then execute an interactive query that runs in the context of the variables observed by TensorWatch and get the returned flow. You can then visualize, save, or process these flows. The developer gave an example of writing an lambda expression that calculates the average weight gradient for each layer in the model at the completion of each batch and sends the result as a tensor stream that can be plotted as a bar graph.

TensorWatch also provides pre-training and post-training features based on existing open source libraries, which are used to provide model graph visualization, data exploration through dimensionality reduction, model statistics and several prediction interpreters for convolution networks.

TensorWatch supports Python 3.x and is tested with PyTorch 0.4-1.x.

Installation:

Linuxidc@linuxidc:~/www.linuxidc.com$ pip3 install tensorwatch

This is the end of the article on "how to install TensorWatch". Thank you for reading! I believe you all have a certain understanding of the knowledge of "how to install TensorWatch". If you want to learn more, you are welcome to follow the industry information channel.

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.