update

README.md

@@ -76,6 +76,22 @@ To evaluate an OOD detection method, you can use the following command:
 `python eval_ood_detection.py --in-dataset {in-distribution dataset} --name {model name} --method {scoring function} [--adv or --corrupt or --adv-corrupt]`
 
 The argument `--adv` is to evaluate L_inf OOD, `--corrupt` is to evaluate Corrupted OOD, `--adv-corrupt` is to evaluate Comp. OOD. You can only specify one of them. Without them, it will evaluate natural OOD. 
+`{in-distribution dataset}` can be `SVHN`, `CIFAR-10` or `CIFAR-100`. 
+`{scoring function}` can be `msp`, `odin`, `mahalanobis`, `sofl`, `rowl`, `ntom` or `atom`. 
+`{model name}` is the name of the model that you have trained. It should match the corresponding scoring function. See the following table for the matchings: 
+
+| model name  | scoring function |
+| ------------- | ------------- |
+| vanilla  |  msp |
+| vanilla  | odin  |
+| vanilla  | mahalanobis  |
+| SOFL     |  sofl |
+| OE       |  msp  |
+| ACET     |  msp |
+| CCU      |  msp |
+| ROWL     |  rowl |
+| NTOM     | ntom |
+| ATOM     | atom | 
 
 ### Citation 
 Please cite our work if you use the codebase: 

eval_ood_detection.py

@@ -25,7 +25,7 @@
 parser = argparse.ArgumentParser(description='Pytorch Detecting Out-of-distribution examples in neural networks')
 
 parser.add_argument('--in-dataset', default="CIFAR-10", type=str, help='in-distribution dataset')
-parser.add_argument('--name', required=True, type=str, help='model name')
+parser.add_argument('--name', required=True, type=str, help='the name of the model trained')
 parser.add_argument('--model-arch', default='densenet', type=str, help='model architecture')
 
 parser.add_argument('--gpu', default = '0', type = str, help='gpu index')