-
## 一言でいうと
従来の畳み込み層の代替えになりうる摂動層を提案。摂動層は畳み込み層よりも少ないパラメータで学習が可能。また、MNIST、CIFAR-10、PASCAL VOCおよびImageNetにおいてCNNと同等の性能を発揮する。
![pnn_1](https://user-images.githubusercontent.com/7046610/41200984-a006e05a-6…
-
### Implement the Guides on the homepage and in search
```[tasklist]
### Coordinate with the server to integrate the new guide (+News?) experience
- [ ] Define a backend to serve the guides (Wo…
-
https://cmsweb.cern.ch/crabserver/ui/task/191214_172243%3Amhadley_crab_Upsilon_nomMass_14Dec2019_Step1
-
get_criteo_feature.py中的特征下标应该是从1开始的,PNN.py中直接使用idx,这样的话使用tf.nn.embedding_lookup进行embedding是从0开始?
-
Hello Kuy
I was wondering if you succeeded to get the DDJ-T1 working with MIXXX?
I did not got it running well
I hope you can help me out.I'm a new bee on githug so sorry if this is not the rig…
-
Hi,
I am trying to train a PHC with my data. But I don't want to train a brand new PHC but based on a pre-trained PHC. The following are the training steps designed based on my understanding:
1. I…
-
I have noticed that there is no way currently to use the clustering statistics in `scale_cuts.py`, even though the `twopoint_wrapper.py` and `twopoint.py` that are used allow for the `w`, `Pnn` and co…
-
I think that in the PNN.py when the k-fold cross-validation method is used, the network must be reinitialized for each cross-validation to create fresh w1 and w2.
The problem now is that the test sam…
-
Moving the workshoppers under the workshopper org umbrella means that there is no longer a signal of who the implicit "lead maintainer" (i.e., the original author) is. Following the example of the oth…
-
As a user backing up a big blog, that goes back a few years, I find that I end up with a directory that chokes `ls`, Finder, and Windows Explorer. I'd like to be able to easily browse the backup with …