Allow the multiprocessing start method to have already been set

src/qusi/internal/infer_session.py

@@ -29,7 +29,11 @@ def infer_session(
     :param device: The device to run the model on.
     :return: A list of arrays with each element being the array predicted for each light curve in the dataset.
     """
-    torch.multiprocessing.set_start_method("spawn")
+    logger.info(f'Creating dataloader workers...')
+    try:
+        torch.multiprocessing.set_start_method("spawn")
+    except RuntimeError:
+        pass
     if workers_per_dataloader == 0:
         prefetch_factor = None
     else:
@@ -41,6 +45,7 @@ def infer_session(
         infer_dataloaders.append(infer_dataloader)
     model.eval()
     results = []
+    logger.info(f'Entering infer loop...')
     for infer_dataloader in infer_dataloaders:
         result = infer_phase(infer_dataloader, model, device=device)
         results.append(result)
@@ -61,6 +66,6 @@ def infer_phase(dataloader, model: Module, device: Device):
             batches_of_predicted_targets.append(batch_predicted_targets_array)
             batch_count += 1
             processed_count += batch_predicted_targets_array.shape[0]
-            logger.info(f'Processed: {processed_count}')
+            logger.info(f'Processed: {processed_count}.')
     predicted_targets = np.concatenate(batches_of_predicted_targets, axis=0)
     return predicted_targets

src/qusi/internal/train_session.py

@@ -86,7 +86,10 @@ def train_session(
         dir=sessions_directory,
     )
     train_dataset = InterleavedDataset.new(*train_datasets)
-    torch.multiprocessing.set_start_method("spawn")
+    try:
+        torch.multiprocessing.set_start_method("spawn")
+    except RuntimeError:
+        pass
     workers_per_dataloader = system_configuration.preprocessing_processes_per_train_process
     if workers_per_dataloader == 0:
         prefetch_factor = None