update

Author: SerialLain3170

UserHint/README.md

@@ -3,13 +3,13 @@
 ## Summary
 ![here](https://github.com/SerialLain3170/Line-to-Color/blob/master/UserHint/network.png)
 
-- 線画と一緒にカラーヒントを与えて、UNet + Discriminatorで生成
-- 線画だけでなく、VGG16を通した線画特徴量も中間層に与える
+- Taking the line art and color hint as input, we generate colored image with the network architecture based on pix2pix.
+- We give hidden layers not only line art but also features of line art extracted by VGG16.
 
 ## Results
-私の環境で生成した例を以下に示す。
+Result generated by my development environment is below.
 ![here](https://github.com/SerialLain3170/Line-to-Color/blob/master/UserHint/example.png)
 
-- バッチサイズは16
-- 最適化手法はAdam(α=0.0002, β1=0.5)
-- 損失関数としてはAdversarial lossとContent loss。重みはそれぞれ0.001と1.0
+- Batch size: 16
+- Using Adam as optimizer
+- The weight of adversarial loss is 0.001 and the weight of content loss is 1.0