The new version of Visual DL has been upgraded in two ways.
First, Visual DL is optimized to further expand the range of support, not only to simplify the process for developers, but also to meet their different needs for Python versions, Python 3.4, 3.5, and 3.6 Manylinux users can be used without barriers.
At the same time, in order to meet the different operational needs of developers, the new version of Visual DL will San (javasript Framework ) switch to Vue, with Vue documentation system more rich and sound features, developers to use the threshold is further reduced, can be extremely convenient and simple to complete the operation.
As you know, Visual DL helps developers visualize the various parameters of the model training process and the computed data flow graphs in real time to help model trainers better understand, debug, and optimize their models.
The developer only needs to write various parameter data related to the model to the log through the interface provided by Visual DL, and then Visual DL will read the model-related data in the log to display it, including various quantitative metrics during the model training process, various pictures generated by user incoming or intermediate training process, and the computational data flow graphs during the neural network training process.
In terms of specific use, the Visual DL deep learning visualization tool is fully functional and has the advantages of being easy to integrate, easy to use, and supporting most deep learning frameworks. At the same time, it changes the traditional diary-style recording format, making it easy for users to visualize the training process and helping to better control the big picture.
Visual DL currently has 1300+ "stars" and over 200 "forks" on Github.