Add Accelerator API to ensembling tutorial (#3606)

## Motivation
- Enhance distributed tutorial: integrate Accelerator API for devise
selection

- Integrate the Accelerator API to support multiple accelerators

Signed-off-by: jafraustro <[email protected]>
Co-authored-by: Svetlana Karslioglu <[email protected]>

Author: jafraustro

intermediate_source/ensembling.py

@@ -50,7 +50,7 @@ def forward(self, x):
 # minibatch of size 64. Furthermore, lets say we want to combine the predictions
 # from 10 different models.
 
-device = 'cuda'
+device = torch.accelerator.current_accelerator()
 num_models = 10
 
 data = torch.randn(100, 64, 1, 28, 28, device=device)