Відео

Thank you for this amazing research and the video. I read the paper and would like to test out the framework. Is there a Github Repos for this project ?

If we look at the experiment, we can see that it does not exceed 20dB. It is understood that this is due to the use of DPRNN as the base model. I think it would be good to develop it based on Sepformer or another model based on Transformer, rather than the DPRNN model.

Overall, it seems like you've done a great job explaining the contents of the PowerPoint. In the experimental results, there are details on Accuracy, but I was wondering if there were any findings utilizing methods like a manifold to visually inspect whether the features of the classes were effectively differentiated after reducing the characteristics of high-dimensional classes to lower dimensions (2D or 3D). Furthermore, the idea of applying strong and weak augmentation to other tasks to improve generalization also appears to be a good approach. Thank you.

Hello, is there any way I can contact you?

Hello, I really like this demo, can you give me the source code of this demo?

Is English version available??

hello, i can't see the video image clearly

can you send presentation

Can it tell the difference between dreamtalk and speech? or snoring and breathe? thx

Can you provide code for learning

Hello

Does this use 2 models? An object detection that takes the whole image as the input to detect persons and then use an action recognition model on each detected person?

How does D_m can be trained on only few data paired data?

any english subtitles ?