-
Notifications
You must be signed in to change notification settings - Fork 2.8k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Feature Request: Automated Creation Based On PyTorch/Keras Neural Network Structures #124
Comments
There is already raw code for PyTorch This will initially be an easy example with explicit fully connected layers. |
Open
2 tasks
PR #126 focuses on PyTorch. TODOs for subsequent PRs
|
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
This feature request is about an automated creation of visualizations from PyTorch/Keras sequential architectures.
Current State
Neural network structures need to be explicitly created for visualization via PlotNeuralNet with the supported, mainly residual convolutional architecture parts, functions.
Proposed State
Neural network structures are visualized based on a given Python class structure, like PyTorch or Keras modules.
Possible Implementation
Proposing a module parsing Python class that retrieves a "net list"
Initial implementation should focus on the simple feed forward network structures. This can be extended based on more complex net lists.
Unfortunately, structures which are not defined in a feed-forward layout or otherwise very structured layout but created via calls of layers in a forward pass function/method are way harder to implement. Asking for support, here.
The text was updated successfully, but these errors were encountered: