Fix typo in identifier of PerceptualNetworkType

    Previously it was 'medical_resnet50_23datasets', which doesn't correspond to the name in the hubconf.py of Warvito's repo. So either on got the error:
    ```
    ValueError: Unrecognised criterion entered for Adversarial Loss. Must be one in: alex, vgg, squeeze, radimagenet_resnet50, medicalnet_resnet10_23datasets, medical_resnet50_23datasets, resnet50
    ```
    or
    ```
    RuntimeError: Cannot find callable medical_resnet50_23datasets in hubconf
    ```

    Fix the identifier's and string's typo from medical_* to medicalnet_* to conform to Warvito's hubconf.py in the MedicalNet repo.
    Add test case for `network_type=medicalnet_resnet50_23datasets`

Signed-off-by: Fabian Klopfer <fabian.klopfer@ieee.org>

Signed-off-by: Fabian Klopfer <fabian.klopfer@ieee.org>

monai/losses/perceptual.py

@@ -29,7 +29,7 @@ class PercetualNetworkType(StrEnum):
     squeeze = "squeeze"
     radimagenet_resnet50 = "radimagenet_resnet50"
     medicalnet_resnet10_23datasets = "medicalnet_resnet10_23datasets"
-    medical_resnet50_23datasets = "medical_resnet50_23datasets"
+    medicalnet_resnet50_23datasets = "medicalnet_resnet50_23datasets"
     resnet50 = "resnet50"
 
 

tests/test_perceptual_loss.py

@@ -40,6 +40,11 @@
         (2, 1, 64, 64, 64),
         (2, 1, 64, 64, 64),
     ],
+    [
+        {"spatial_dims": 3, "network_type": "medicalnet_resnet50_23datasets", "is_fake_3d": False},
+        (2, 1, 64, 64, 64),
+        (2, 1, 64, 64, 64),
+    ],
     [
         {"spatial_dims": 3, "network_type": "resnet50", "is_fake_3d": True, "pretrained": True, "fake_3d_ratio": 0.2},
         (2, 1, 64, 64, 64),