Discussion of article "Neural networks made easy (Part 23): Building a tool for Transfer Learning"
You are missing trading opportunities:
- Free trading apps
- Over 8,000 signals for copying
- Economic news for exploring financial markets
Registration
Log in
You agree to website policy and terms of use
If you do not have an account, please register
New article Neural networks made easy (Part 23): Building a tool for Transfer Learning has been published:
In this series of articles, we have already mentioned Transfer Learning more than once. However, this was only mentioning. in this article, I suggest filling this gap and taking a closer look at Transfer Learning.
So, I see three clear blocks. In the first block we will work with the donor model. Here we need the ability to select a file with a trained model. After loading a model from a file, the tool must provide a description of the architecture of the loaded model. This is because the user should understand which model is loaded and which neural layers it will copy. We will also inform the tool about the number of copied layers. As mentioned above, we will sequentially copy the neural layers starting from the source data layer.
In the second block, neural layers will be added. Here we will create fields for entering information about the neural layer being created. As with the program code, we will sequentially describe each neural layer one by one and will add it to the architecture of the new model.
The third block will display the holistic architecture of the created model with the ability to specify a file to save it. An example design of the tool is presented below.
Both the design of the tool and its implementation are presented for demonstration purposes only. You can always change them to best meet your needs.
Author: Dmitriy Gizlyk