-
As described in your paper, the Gumbel Softmax trick is applied to enable the back-propagation. The gradients must go through the sampled adjacency matrix to the edge predictor. However, your implemen…
-
> Most frontends should be implemented in very few lines of code, with many only being a single line of code. Shortening a frontend might involve adding a new function to our Experimental API, or exte…
-
https://arxiv.org/abs/2002.12416
RGB input 활용이 아니라, YCrCb 를 DCT 후 input으로 씀
더 큰 이미지를 network input으로 활용할 수 있어서 정확도 향상에 좋음
![image](https://user-images.githubusercontent.com/16400591/78205422-05a9…
-
Really interesting work! I'm trying to test the effectiveness of PRIOR's report reconstruction on some clinical data, and I'm wondering if there's an efficient way to run inference without making sign…
-
Hi, what does the latent_dim mean in your code? Could it be changed to other numbers? I can understand that categorical_dim means 10 categories for 10 digits, but I'm confused about the latent_dim. Th…
-
Hi compvis group, thanks for your overwheming work and i'd love to deploy your method into my project.
However, i found there is no learning rate scheduler(lr scheduler) was used for training.
Mo…
-
Hi
As per the paper, 4 features must be concatenated as input to TRUNET,
1. log spectrum
2. PCEN
3. real part of demodulated phase
4. imaginary part of demodulated phase
so the input will…
-
作者您好,我理解deit_dge.py中,应该是使用了 Dynamic Grained Encoder。但是在代码dge.py中,sparse_train默认是false,并且这个参数在DynamicGrainedEncoder初始化的时候也没改,那么训练的时候走的是这个循环吗?
https://github.com/StevenGrove/vtpack/blob/38d5af0c3d91811…
-
May I know why [this training code](https://colab.research.google.com/drive/1v5wY22CkyvKPz21tdwSMPv0T3fsIro0D?usp=sharing#scrollTo=6qJRPd9-sEdK) still gives CUDA-out-of-memory issue even after DeepSpe…
-
Thank you for the great work!
I tried to repoduce `VQGAN OpenImages (f=8), 8192, GumbelQuantization` model based on the config file from the cloud. (the detailed config file is in below.)
VQGA…