Minor fixes to remove bug when running with no valid set.

kausmees · Jun 23, 2021 · 969353e · 969353e
1 parent 5d5fe12
commit 969353e
Show file tree

Hide file tree

Showing 2 changed files with 15 additions and 3 deletions.
diff --git a/run_gcae.py b/run_gcae.py
@@ -670,7 +670,10 @@ def loss_func(y_pred, y_true):
 		print("Model layers and dimensions:")
 		print("-----------------------------")
 
-		output_valid_batch, encoded_data_valid_batch = autoencoder(input_valid[0:2], is_training = False, verbose = True)
+		input_test, targets_test, _  = dg.get_train_set(0.0)
+		if not missing_mask_input:
+			input_test = input_test[:,:,0, np.newaxis]
+		output_test, encoded_data_test = autoencoder(input_test[0:2], is_training = False, verbose = True)
 
 		######### Create objects for tensorboard summary ###############################
 

diff --git a/utils/data_handler.py b/utils/data_handler.py
@@ -78,7 +78,7 @@ def _define_samples(self):
 		self.n_train_samples_orig = len(self.sample_idx_all)
 		self.n_train_samples = self.n_train_samples_orig
 		self.ind_pop_list_train_orig = ind_pop_list[self.sample_idx_all]
-		self.train_set_indices = np.array(range(self.n_train_samples))
+		self.train_set_indices = np.arange(self.n_train_samples)
 
 		self.n_valid_samples = 0
 
@@ -170,7 +170,10 @@ def define_validation_set(self, validation_split):
 
 		_, _, self.sample_idx_train, self.sample_idx_valid = get_test_samples_stratified(self.genotypes_train_orig, self.ind_pop_list_train_orig, validation_split)
 
-		self.train_set_indices = np.array(range((len(self.sample_idx_train))))
+		self.sample_idx_train = np.array(self.sample_idx_train)
+		self.sample_idx_valid = np.array(self.sample_idx_valid)
+
+		self.train_set_indices = np.array(range(len(self.sample_idx_train)))
 		self.n_valid_samples = len(self.sample_idx_valid)
 		self.n_train_samples = len(self.sample_idx_train)
 
@@ -184,8 +187,14 @@ def get_valid_set(self, sparsify):
 				 target_data_valid (n_valid_samples x n_markers): original validation genotypes
 				 ind_pop_list valid (n_valid_samples x 2) : individual and population IDs of validation samples
 
+				 or
+				 empty arrays if no valid set defined
+
 		'''
 
+		if self.n_valid_samples == 0:
+			return np.array([]), np.array([]), np.array([])
+
 		# n_valid_samples x n_markers x 2
 		input_data_valid = np.full((len(self.sample_idx_valid),self.genotypes_train_orig.shape[1],2),1.0)