Skip to content
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

Support residual/skip connections #56

Open
dario-loi opened this issue Apr 19, 2023 · 2 comments
Open

Support residual/skip connections #56

dario-loi opened this issue Apr 19, 2023 · 2 comments

Comments

@dario-loi
Copy link

dario-loi commented Apr 19, 2023

Not an issue, just a simple suggestion, could it be possible to introduce residual/skip connections in the form of extra directed arrows in the network graph? perhaps the skip connection could feed into a node where the aggregation operator can be specified in LaTeX (Addition, or any other function).

I understand that skip connections might be complext on the UX side since they kind of break the "linearity" of the list of layers, however, having them would allow a complete modeling of almost all possible architectures using CNNs, so I'd really appreciate the addition.

@alexlenail
Copy link
Owner

In general, I've been trying to avoid adding support for network layouts that are best specified by code, because tensorboard model graphs already exists. The only diagrams I've seen of resnets so far look like those that tensorboard would support:

Doing some googling now, I saw this diagram which was neat:

What did you have in mind?

@dario-loi
Copy link
Author

The second one sure does look very neat! When I originally wrote the post I had something like this in mind:

network

Which is a very simplistic diagram of an EDSR-inspired network that I spun op on draw.io in a dozen minutes or so (spoiler, the architecture doesn't work).

So not necessarily a very dense network like RDN, just a single skip connection would suffice.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

2 participants