dis v2

IS-Net/train_valid_inference_main.py

@@ -709,9 +709,9 @@ if __name__ == "__main__":
     print("building model...")
     hypar["model"] = ISNetDIS() #U2NETFASTFEATURESUP()
     hypar["early_stop"] = 20 ## stop the training when no improvement in the past 20 validation periods, smaller numbers can be used here e.g., 5 or 10.
-    hypar["model_save_fre"] = 2000 ## valid and save model weights every 2000 iterations
+    hypar["model_save_fre"] = 20 ## valid and save model weights every 2000 iterations
 
-    hypar["batch_size_train"] = 8 ## batch size for training
+    hypar["batch_size_train"] = 2 ## batch size for training
     hypar["batch_size_valid"] = 1 ## batch size for validation and inferencing
     print("batch size: ", hypar["batch_size_train"])
 

README.md

@@ -25,7 +25,9 @@
 
 <br> 
 
-## 1. [Our DIS5K Dataset V1.0 (Version Alias: DIS5K Sailship)](https://xuebinqin.github.io/dis/index.html) 
+## 1. Our Dichotomous Image Segmentation (DIS) Dataset
+
+### 1.1 [DIS dataset V1.0: DIS5K](https://xuebinqin.github.io/dis/index.html) 
 
 <br>
 
@@ -35,6 +37,15 @@
 ![complexities-qual](figures/complexities-qual.jpeg)
 ![categories](figures/categories.jpeg)
 
+### 1.2 [DIS dataset V2.0](https://github.com/xuebinqin/DIS)
+
+<br>
+
+Although our DIS5K V1.0 includes samples from more than 200 categories, many categories, such as human, animals, cars and so on, in real world are not included. [So the current version (v1.0) of our dataset may limit the robustness of the trained models.]() To build the comprehensive and large-scale highly accurate dichotomous image segmentation dataset, we are building our DIS dataset V2.0. The V2.0 will be released soon. Please stay tuned.
+
+Samples from DIS dataset V2.0.
+![dis-v2](figures/dis-v2.jpg)
+
 <br> 
 
 ## 2. APPLICATIONS of Our DIS5K Dataset