[ADD] Missing Batchnorm (#317)

* Add batch norm to mlp * Increase max runtime in test * Try drop last for running batch norm * Change order to BN-ReLU

Author: ravinkohli

autoPyTorch/pipeline/components/setup/network_backbone/MLPBackbone.py

@@ -55,6 +55,7 @@ def _add_layer(self, layers: List[nn.Module], in_features: int, out_features: in
 
  """
  layers.append(nn.Linear(in_features, out_features))
+ layers.append(nn.BatchNorm1d(out_features))
  layers.append(_activations[self.config["activation"]]())
  if self.config['use_dropout']:
  layers.append(nn.Dropout(self.config["dropout_%d" % layer_id]))

autoPyTorch/pipeline/components/setup/network_backbone/ShapedMLPBackbone.py

@@ -62,6 +62,7 @@ def _add_layer(self, layers: List[nn.Module],
  in_features: int, out_features: int, dropout: float
  ) -> None:
  layers.append(nn.Linear(in_features, out_features))
+ layers.append(nn.BatchNorm1d(out_features))
  layers.append(_activations[self.config["activation"]]())
  if self.config["use_dropout"] and self.config["max_dropout"] > 0.05:
  layers.append(nn.Dropout(dropout))

autoPyTorch/pipeline/components/training/data_loader/base_data_loader.py

@@ -114,7 +114,7 @@ def fit(self, X: Dict[str, Any], y: Any = None) -> torch.utils.data.DataLoader:
  shuffle=True,
  num_workers=X.get('num_workers', 0),
  pin_memory=X.get('pin_memory', True),
- drop_last=X.get('drop_last', False),
+ drop_last=X.get('drop_last', True),
  collate_fn=custom_collate_fn,
  )
 

test/test_pipeline/test_tabular_classification.py

@@ -460,7 +460,7 @@ def test_train_pipeline_with_runtime(fit_dictionary_tabular_dummy):
  # Convert the training to runtime
  fit_dictionary_tabular_dummy.pop('epochs', None)
  fit_dictionary_tabular_dummy['budget_type'] = 'runtime'
- fit_dictionary_tabular_dummy['runtime'] = 3
+ fit_dictionary_tabular_dummy['runtime'] = 5
  fit_dictionary_tabular_dummy['early_stopping'] = -1
 
  pipeline = TabularClassificationPipeline(
@@ -474,11 +474,11 @@ def test_train_pipeline_with_runtime(fit_dictionary_tabular_dummy):
  run_summary = pipeline.named_steps['trainer'].run_summary
  budget_tracker = pipeline.named_steps['trainer'].budget_tracker
  assert budget_tracker.budget_type == 'runtime'
- assert budget_tracker.max_runtime == 3
+ assert budget_tracker.max_runtime == 5
  assert budget_tracker.is_max_time_reached()
 
  # There is no epoch limitation
  assert not budget_tracker.is_max_epoch_reached(epoch=np.inf)
 
- # More than 200 epochs would have pass in 3 seconds for this dataset
+ # More than 200 epochs would have pass in 5 seconds for this dataset
  assert len(run_summary.performance_tracker['start_time']) > 100