-
Dear David Ha, dear Jürgen Schmidhuber
Thank for this inspirational blog-post. I have stumbled upon your paper while researching for my BSc thesis. It is concerned with training agents to navigat…
-
Why is using SparseUNet (SpUNet v1m1) better than PTv2 (ptv2m2) when training my dataset?
The val mＩoＵ of SpUNet－v1m1 is 0.873 , while ptv2m2 is 0.7261．
, the "`train_blocks must be single for split mode`" error suddenly occurred. This error had not appea…
-
Description: The current implementation of the check flag feature does not provide a mechanism to display only failing checks, specific checks of interest, or exclude certain checks from the output. T…
-
## 一言でいうと
通常学習後に観測するActivation Mapを、Attentionとしてネットワーク内に組み込んだ研究。Activation Mapの計算には特徴マップ以外にクラス分類への貢献を測る重みが必要だが(通常は全結合層の重みを使う)、これを取得するためAttention側からもクラス分類確率を出力し、マルチタスクで学習している。
![image](https://us…
-
## Description:
Hello! I’ve been following the development of this repository and appreciate the efforts to benchmark various efficient Transformer variants. I’d like to propose the implementation of…