Demonstrate the neural network structure of each experimental

[little1d]

🤩 Functional description

Being new to deep learning and even pytorch, I have a hard time understanding the intricate network structure in deep learning. What are the roles of the convolutional, pooling and Softmax layers? How are they connected? A visual representation would improve development and learning efficiency.

I need a visual structure to help me understand urgently!!!.

👍 What problem does this feature solve

1. Each experiment dashboard displays a different diagram of the network architecture to make it easier to compare experiments.
2. The model building process is easier to understand.
3. Make model fine-tuning process more convenient.

👾 What does the proposed application interface look like?

Maybe swanlab.struct? You can add this code under model define process, take this picture as example:

🚑 Any additional information [e.g. screenshots].

This is the current work in several related fields, I hope it will be helpful to you.

https://github.com/HarisIqbal88/PlotNeuralNet

https://github.com/julrog/nn_vis