update read me

Author: wbw520

README.md

@@ -41,3 +41,22 @@ Then see the generated concepts by:
 python vis_retri.py --num_classes 50 --num_cpt 20 --base_model resnet18 --index 300 --top_sample 20 --dataset CUB200
 ```
 
+#### Demo weight
+We provide some trained weight for demo. Download and set them to folder "saved_model/"
+![MNIST](https://drive.google.com/file/d/1wQtsi2jTEoG1k877XNG9cB4njWztlkzn/view?usp=sharing)
+![CUB 50 classes](https://drive.google.com/file/d/1XIcTPCCb3uXFOOb_PrNXjJHfMrh--wsy/view?usp=sharing)
+![ImageNet 50 classes](https://drive.google.com/file/d/1VSAlC6QftQDUzIAE8WJ6D0tNEaa1niLR/view?usp=sharing)
+
+Run them as following:
+```
+MNIST
+python vis_recon.py --num_classes 10 --num_cpt 20 --index 0 --top_sample 20 --top_sample 20 --deactivate -1
+
+CUB
+python process.py
+python vis_retri.py --num_classes 50 --num_cpt 20 --base_model resnet18 --index 300 --top_sample 20 --dataset CUB200
+
+ImageNet
+python process.py
+python vis_retri.py --num_classes 50 --num_cpt 20 --base_model resnet18 --index 300 --top_sample 20 --dataset CUB200
+```

main_recon.py

@@ -9,6 +9,9 @@
 import os
 
 
+os.makedirs('saved_model/', exist_ok=True)
+
+
 def main():
     transform = transforms.Compose([transforms.ToTensor(), transforms.Normalize((0.1307,), (0.3081,))])
     trainset = datasets.MNIST('../data', train=True, download=True, transform=transform)

main_retri.py

@@ -8,6 +8,9 @@
 from utils.tools import fix_parameter, print_param
 
 
+os.makedirs('saved_model/', exist_ok=True)
+
+
 def main():
     model = MainModel(args)
     device = torch.device(args.device)