Is your feature request related to a problem? Please describe.
I would like to implement some kind of Deepfake detector inside this codebase. I'm not sure where to start. It could be a neural net under pyvisim/neural_networks (e.g., a net that extracts CLIP features and use it to give a deepfake score). But that's just a suggestion, there might be other ways.
The final goal is to have a class that assigns a score from [0-1] that tells how "fake" an image is.
Describe the solution you'd like
We can consider implementing a defined algorithm (ML-based, of course). Feel free to suggest any cool algorithm you found on papers. If it's a neural network, it would be placed at pyvisim/neural_networks.
Btw, I found a repo that summarizes these papers:
Describe alternatives you've considered
I'm not sure yet, feel free to share your idea :)
Additional context
Feel free to get yourself inspiration from these implementations:
Is your feature request related to a problem? Please describe.
I would like to implement some kind of Deepfake detector inside this codebase. I'm not sure where to start. It could be a neural net under
pyvisim/neural_networks(e.g., a net that extracts CLIP features and use it to give a deepfake score). But that's just a suggestion, there might be other ways.The final goal is to have a class that assigns a score from
[0-1]that tells how "fake" an image is.Describe the solution you'd like
We can consider implementing a defined algorithm (ML-based, of course). Feel free to suggest any cool algorithm you found on papers. If it's a neural network, it would be placed at
pyvisim/neural_networks.Btw, I found a repo that summarizes these papers:
Describe alternatives you've considered
I'm not sure yet, feel free to share your idea :)
Additional context
Feel free to get yourself inspiration from these implementations: