pre.load_data(filename='test_public_norm.csv', name='validate')
X_train_df = pre.get(name='train').drop(columns=['emotion'])
y_train_df = pre.get(name='train')['emotion']
X_val_df = pre.get(name='validate').drop(columns=['emotion'])
y_val_df = pre.get(name='validate')['emotion']
dtype = torch.float
model_name = f'cnn_triple_layer_D_bs_{learning_rate}_{batch_size}_{n_epochs}_{n_classes}'
model = CnnTripleLayer(model_name, d_out=n_classes)
model.train()
train_classifier = TrainClassifier2(model,
X_train_df,
y_train_df,
X_val_df,
y_val_df,
root_dir=current_working_dir)
t = time.time()
trained_model, optimizer, criterion, \
train_loss_hist, train_acc_hist, train_f1_hist, train_b_hist,\
val_loss_hist, val_acc_hist, val_f1_hist, val_b_hist = train_classifier.run_train(n_epochs=n_epochs,
lr=learning_rate,
batch_size=batch_size)
print(f'trained in {time.time() - t} sec')
if args.s_model:
m_exporter = ModelExporter('fer2013_datasetD',
root_dir=current_working_dir)
m_exporter.save_nn_model(trained_model, optimizer,
trained_model.get_args())
y_val_df = pre.get(name='val')['emotion']
n_classes = 7
n_epochs = 10
learning_rate = 0.001
batch_size = 64
dtype = torch.float
model_conv = torchvision.models.vgg16_bn(pretrained=True) # _bn batch normalization
model_name = f'cnn_VGG16_pretrained_A_bs_{learning_rate}_{batch_size}_{n_epochs}_{n_classes}'
model = CnnVGG16Pretrained(model_name, features = model_conv.features)
model.load_state_dict(state_dict=model_conv.state_dict())
model.prepare_model()
model.train()
train_classifier = TrainClassifier2(model, train_pixel_224_np, y_train_df, val_pixel_224_np, y_val_df, to_rgb=True,
root_dir=current_working_dir)
t = time.time()
trained_model, optimizer, criterion, \
train_loss_hist, train_acc_hist, train_f1_hist, train_b_hist,\
val_loss_hist, val_acc_hist, val_f1_hist, val_b_hist = train_classifier.run_train(n_epochs=n_epochs,
lr=learning_rate,
batch_size=batch_size)
print(f'trained in {time.time() - t} sec')
if args.s_model:
m_exporter = ModelExporter('fer2013_datasetA', root_dir=current_working_dir)
m_exporter.save_nn_model(trained_model, optimizer,trained_model.get_args())
m_exporter.save_results(f'{model_name}',
train_loss_hist, train_acc_hist, train_f1_hist, train_b_hist,
val_loss_hist, val_acc_hist, val_f1_hist, val_b_hist)
batch_size = 32
pre = Preprocessing('fer2013')
pre.load_data(filename='DatasetD.csv', name='train')
X_df = pre.get(name='train').drop(columns=['emotion'])
y_df = pre.get(name='train')['emotion']
dtype = torch.float
device = torch.device("cpu")
model_name = f'cnn_double_layer_D_bs_{learning_rate}_{batch_size}_{n_epochs}_{n_classes}'
model = CnnDoubleLayer(model_name, d_out=n_classes)
model.train()
train_classifier = TrainClassifier2(model, X_df, y_df)
t = time.time()
trained_model, optimizer, criterion, loss_hist, loss_val_hist, f1_val_hist = train_classifier.run_train(
n_epochs=n_epochs, lr=learning_rate, batch_size=batch_size)
print(f'trained in {time.time() - t} sec')
pre.save_results(loss_hist, loss_val_hist, f1_val_hist, f'{model_name}')
if args.s_model:
m_exporter = ModelExporter('fer2013_DatasetD')
m_exporter.save_nn_model(trained_model, optimizer,
trained_model.get_args())
if args.s_patterns:
detected_patterns1 = trained_model.get_detected_patterns1()
for idx in range(10):
plt.figure(1, figsize=(20, 10))
X_train_df = pre.get(name='train').drop(columns=['emotion'])
y_train_df = pre.get(name='train')['emotion']
X_val_df = pre.get(name='validate').drop(columns=['emotion'])
y_val_df = pre.get(name='validate')['emotion']
n_classes = 7
n_epochs = 100
learning_rate = 0.001
batch_size = 64
model_name = f'cnn_simple_reduced_bs_{learning_rate}_{batch_size}_{n_epochs}_{n_classes}'
model = CnnSimple(model_name, d_out=n_classes)
train_classifier = TrainClassifier2(model,
X_train_df,
y_train_df,
X_val_df,
y_val_df,
root_dir=script_root_dir)
t = time.time()
trained_model, optimizer, criterion, \
train_loss_hist, train_acc_hist, train_f1_hist, train_b_hist,\
val_loss_hist, val_acc_hist, val_f1_hist, val_b_hist = train_classifier.run_train(n_epochs=n_epochs,
lr=learning_rate,
batch_size=batch_size)
print(f'trained in {time.time() - t} sec')
if args.s_model:
m_exporter = ModelExporter('fer2013_reduced', root_dir=script_root_dir)
m_exporter.save_nn_model(trained_model, optimizer,
trained_model.get_args())