-
## 一言でいうと
High Wayの手法をRNNに適用する話。High Wayは入力をバイパスするゲートCを設けて、これと隠れ層HをゲートTに通したものを合算させることで入力にない表現のみ学習をさせるような手法。これで伝搬ステップの深いRNNを作る。言語モデル(PTB)とWikipediaの語予測でSOTA
### 論文リンク
https://arxiv.org/abs/160…
-
A highway LSTM ([Srivastava et al., 2015](https://papers.nips.cc/paper/5850-training-very-deep-networks.pdf); [Zhang et al., 2016](https://groups.csail.mit.edu/sls/publications/2016/YuZhang3_ICASSP-16…
-
[Deep Learning Powered In-Session Contextual Ranking using Clickthrough Data](http://ftp.cs.wisc.edu/machine-learning/shavlik-group/li.nips14.pdf)
[Techniques for Deep Query Understanding ](https://a…
-
Post a reading of your own that uses deep learning for social science analysis and understanding, with a focus on network, graph, or tabular data.
-
As we move from images to videos, it seems imperative to feed sequential data into common image layers. However, the problem of dealing sequential data with such layers is not clears on Pytorch. I sug…
-
@benanne @dnouri @craffel @f0k @skaae (and anyone else of course)
With first release imminent it would be nice to have a bit more here... I know a couple of you have stuff written up already, but I b…
-
### Description of the bug:
Hi there,
I am trying to convert my LSTM custom model with the library to put it on M5Stack ESP32 device.
I occur in an error during the conversion though. I share m…
-
## Keyword: super resolution
There is no result
## Keyword: gan
### Towards Discovery and Attribution of Open-world GAN Generated Images
- **Authors:** Sharath Girish, Saksham Suri, Saketh Rambhatla…
-
PyText: A Seamless Path from NLP research to production
Ahmed Aly, Kushal Lakhotia, Shicong Zhao, Mrinal Mohit, Barlas Oguz, Abhinav Arora, Sonal Gupta, Christopher Dewan, Stef Nelson-Lindall, Rushin…
-
Has anyone tried downscaling the K and/or Q matrices for repeated layers in franken-merges? This should act like changing the temperature of the softmax and effectively smooth the distribution:
**H…