def test_initialisation(self): module = SimpleMLP() module = module.to(self.device) old_classifier_weight = torch.clone(module.classifier.weight) old_classifier_bias = torch.clone(module.classifier.bias) module = as_multitask(module, "classifier") module = module.to(self.device) new_classifier_weight = torch.clone( module.classifier.classifiers["0"].classifier.weight) new_classifier_bias = torch.clone( module.classifier.classifiers["0"].classifier.bias) self.assertTrue( torch.equal(old_classifier_weight, new_classifier_weight)) self.assertTrue(torch.equal(old_classifier_bias, new_classifier_bias))
final_train_data_x.append(temp_train_data_x) final_train_data_y.append(temp_train_data_y) final_test_data_x.append(temp_test_data_x) final_test_data_y.append(temp_test_data_y) final_train_data_y = [x.long() for x in final_train_data_y] final_test_data_y = [x.long() for x in final_test_data_y] return final_train_data_x, final_train_data_y, final_test_data_x, final_test_data_y # MODEL CREATION device = torch.device("cuda:0" if torch.cuda.is_available() else "cpu") model = SimpleMLP(num_classes=2, input_size=38, hidden_size=100) model.to(device) perm = { "1": [13, 22, 20, 14, 6], "2": [9, 10, 0, 1, 2], "3": [11, 15, 17, 21], "4": [18, 19, 7, 8, 12], "5": [3, 4, 5, 16], } task_order_list = [perm] dataset = task_ordering(task_order_list[0]) generic_scenario = tensor_scenario( train_data_x=dataset[0],