Merge pull request #1818 from usamaimdadsian/tensorrt-batch-features-…

…issue

Fix: Mistake in handling the output features of ReID model in TensorRT

boxmot/appearance/backends/tensorrt_backend.py

@@ -116,7 +116,7 @@ def forward(self, im_batch):
             # Execute inference
             self.context.execute_v2(list(self.binding_addrs.values()))
             features = self.bindings["output"].data
-            resultant_features.append(features)
+            resultant_features.append(features.clone())
 
         if len(resultant_features)== 1:
             return resultant_features[0]