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

For the image style transfer task #15

Closed
westream opened this issue Jun 9, 2020 · 2 comments
Closed

For the image style transfer task #15

westream opened this issue Jun 9, 2020 · 2 comments

Comments

@westream
Copy link

westream commented Jun 9, 2020

No description provided.

@westream
Copy link
Author

westream commented Jun 9, 2020

Hello, your work is very inspiring to me. For the image style transfer task, since the image feature map has been down-sampled, the channel domain information becomes very important. I would like to ask about this method of you, how to use my own trained model to visualize under your framework, or you Have any visualizations done similar tasks? For the convolutional network of this generation task, what knowledge does the convolutional nerve learn. thank you very much

@westream westream changed the title Hello, your work is very inspiring to me. For the image style transfer task, since the image feature map has been down-sampled, the channel domain information becomes very important. I would like to ask about this method of you, how to use my own trained model to visualize under your framework, or you Have any visualizations done similar tasks? For the convolutional network of this generation task, what knowledge does the convolutional nerve learn. thank you very much For the image style transfer task Jun 9, 2020
@xiaohk
Copy link
Member

xiaohk commented Jun 9, 2020

Thanks for your interests! This issue is related to #2 and #8.

Currently CNN Explainer only supports the Tiny-VGG architecture that we described in our manuscript. If you want to use a different CNN model, then you would need to modify the code. Here are some functions that you would need to change: #8 (comment).

For GANs model, you can also check out our other work: GAN Lab. Both CNN Explainer and GAN Lab aim to help beginners study the mathematical and algorithmic concepts of CNNs and GANs. They are not particularly designed for domain experts to interpret what each neuron has learned. For interpretation tasks, you can check out Summit and some other work.

I will close this issue for now. Feel free to reply or reopen it if you need further help :)

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