update mmar tests (#4091)

* update mmar tests

Signed-off-by: Wenqi Li <[email protected]>

* [pre-commit.ci] auto fixes from pre-commit.com hooks

for more information, see https://pre-commit.ci

* fixes pylint error

Signed-off-by: Wenqi Li <[email protected]>

Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com>

Author: wyli

tests/ngc_mmar_loading.py

@@ -10,13 +10,15 @@
 # limitations under the License.
 
 import os
+import sys
 import unittest
 
 import torch
 from parameterized import parameterized
 
 from monai.apps.mmars import MODEL_DESC, load_from_mmar
 from monai.config import print_debug_info
+from monai.networks.utils import copy_model_state
 
 
 class TestAllDownloadingMMAR(unittest.TestCase):
@@ -26,10 +28,33 @@ def setUp(self):
 
     @parameterized.expand((item,) for item in MODEL_DESC)
     def test_loading_mmar(self, item):
+        if item["name"] == "clara_pt_self_supervised_learning_segmentation":  # test the byow model
+            default_model_file = os.path.join("ssl_models_2gpu", "best_metric_model.pt")
+            pretrained_weights = load_from_mmar(
+                item=item["name"],
+                mmar_dir="./",
+                map_location="cpu",
+                api=True,
+                model_file=default_model_file,
+                weights_only=True,
+            )
+            pretrained_weights = {k.split(".", 1)[1]: v for k, v in pretrained_weights["state_dict"].items()}
+            sys.path.append(os.path.join(f"{item['name']}", "custom"))  # custom model folder
+            from vit_network import ViTAutoEnc  # pylint: disable=E0401
+
+            model = ViTAutoEnc(
+                in_channels=1,
+                img_size=(96, 96, 96),
+                patch_size=(16, 16, 16),
+                pos_embed="conv",
+                hidden_size=768,
+                mlp_dim=3072,
+            )
+            _, loaded, not_loaded = copy_model_state(model, pretrained_weights)
+            self.assertTrue(len(loaded) > 0 and len(not_loaded) == 0)
+            return
         if item["name"] == "clara_pt_fed_learning_brain_tumor_mri_segmentation":
             default_model_file = os.path.join("models", "server", "best_FL_global_model.pt")
-        elif item["name"] == "clara_pt_self_supervised_learning_segmentation":
-            default_model_file = os.path.join("models_2gpu", "best_metric_model.pt")
         else:
             default_model_file = None
         pretrained_model = load_from_mmar(