-
-
See https://github.com/Mikolaj/horde-ad/issues/9#issuecomment-1086211063.
To test the bindings, implement convolution in horde-ad tests using them.
-
## 一言でいうと
機械学習でプログラムのエラー修正を行う研究。プログラムの行とエラーメッセージそれぞれをBi-LSTMでEncode=>Graph Attention=>LSTM=>各行の表現にエラーの表現を結合し発生位置予測&修正性を行う。稼働可能なコードを編集し、発生したエラーで学習データを生成している。
 using LSTM? As there are some open sou…
-
## 論文リンク
https://cdn.openai.com/research-covers/language-unsupervised/language_understanding_paper.pdf
## 公開日（yyyy/mm/dd）
2018/06/11
## 概要
いわゆる GPT-1 の論文。
言語モデルの教師なし事前学習がその後の教師あり fine-tuning…
-
Abstract: We present Compositional Attention Networks, a novel fully differentiable neural network architecture, designed to facilitate explicit and expressive reasoning. While many types of neural ne…
-
We now have the reasonable capacity to perform time series forecasting via reduction to regression. Some bespoke forecasting architectures would also be good to implement too, though.
@mloning thi…
-
Hello. Let me ask you a question.
I tried to build a BERT-LSTM model for your paper using Movie Reviews data, but I couldn't reproduce the paper results.
My results are as follows.
Training: trai…
-
Hello,
our GPU requires CUDA 11 and therefore I am stuck with using TensoFlow 2.x. I was trying to lift the code to TF2, but it seems I can't find a replacement for `tf.contrib.rnn.CoupledInputForget…