def objective(trial):
out_channels = int(trial.suggest_discrete_uniform('out_channels', 50, 200, 50))
drop_rate = trial.suggest_discrete_uniform('drop_rate', 0.0, 0.5, 0.1)
learning_rate = trial.suggest_loguniform('learning_rate', 5e-4, 5e-2)
momentum = trial.suggest_discrete_uniform('momentum', 0.5, 0.9, 0.1)
batch_size = int(trial.suggest_discrete_uniform('batch_size', 16, 128, 16))
VOCAB_SIZE = len(set(word_to_id.values())) + 1
PADDING_IDX = len(set(word_to_id.values()))
EMB_SIZE = 300
KERNEL_HEIGHTS = 3
STRIDE = 1
PADDING = 1
OUTPUT_SIZE = 4
CONV_PARAMS = [[2, 0], [3, 1], [4, 2]]
NUM_EPOCHS = 10
model = CNN(VOCAB_SIZE, EMB_SIZE, PADDING_IDX, OUTPUT_SIZE, OUT_CHANNELS, KERNEL_HEIGHTS, STRIDE, PADDING, emb_weights=weights)
criterion = nn.CrossEntropyLoss()
optimizer = torch.optim.SGD(model.parameters(), lr=learning_rate, momentum=momentum)
device = torch.device('cuda')
log = train_model(dataset_train, dataset_dev, batch_size, model, criterion, optimizer, NUM_EPOCHS, collate_fn=Padsequence(PADDING_IDX), device=device)
loss_dev, _ = calculate_loss_and_acc(model, dataset_dev, criterion=criterion, device=device)
test_set = CreateDataset(x_test, y_test, test_text2id)
VOCAB_SIZE = len(set(tr_word2id.values())) + 1
EMB_SIZE = 300
PADDING_IDX = len(set(tr_word2id.values()))
OUTPUT_SIZE = 4
HIDDEN_SIZE = 50
NUM_LAYERS = 2
LEARNING_RATE = 5e-2
BATCH_SIZE = 32
NUM_EPOCHS = 10
model = RNN(VOCAB_SIZE,
EMB_SIZE,
PADDING_IDX,
OUTPUT_SIZE,
HIDDEN_SIZE,
NUM_LAYERS,
bidirectional=True)
criterion = nn.CrossEntropyLoss()
optimizer = torch.optim.SGD(model.parameters(), lr=LEARNING_RATE)
log = train_model(train_set,
valid_set,
BATCH_SIZE,
model,
criterion,
optimizer,
NUM_EPOCHS,
collate_fn=Padsequence(PADDING_IDX))
OUTPUT_SIZE = 4
HIDDEN_SIZE = 50
LEARNING_RATE = 1e-3
BATCH_SIZE = 32
NUM_EPOCHS = 10
rnn = RNN(VOCAB_SIZE, EMB_SIZE, PADDING_IDX, OUTPUT_SIZE, HIDDEN_SIZE)
device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')
criterion = nn.CrossEntropyLoss()
optimizer = torch.optim.SGD(rnn.parameters(), lr=LEARNING_RATE)
log = train_model(dataset_train,
dataset_dev,
BATCH_SIZE,
rnn,
criterion,
optimizer,
NUM_EPOCHS,
collate_fn=Padsequence(PADDING_IDX),
device=device)
"""
epoch: 1, loss_train: 1.0687, accuracy_train: 0.6129, loss_dev: 1.1163, accuracy_dev: 0.5926
epoch: 2, loss_train: 1.0061, accuracy_train: 0.6321, loss_dev: 1.0401, accuracy_dev: 0.6053
epoch: 3, loss_train: 0.9022, accuracy_train: 0.6654, loss_dev: 0.9739, accuracy_dev: 0.6148
epoch: 4, loss_train: 0.8375, accuracy_train: 0.6977, loss_dev: 0.9593, accuracy_dev: 0.6370
epoch: 5, loss_train: 0.8625, accuracy_train: 0.6774, loss_dev: 0.9510, accuracy_dev: 0.6233
epoch: 6, loss_train: 0.7032, accuracy_train: 0.7437, loss_dev: 0.8379, accuracy_dev: 0.6794
epoch: 7, loss_train: 0.6693, accuracy_train: 0.7564, loss_dev: 0.8294, accuracy_dev: 0.6815
epoch: 8, loss_train: 0.5748, accuracy_train: 0.7857, loss_dev: 0.7458, accuracy_dev: 0.7090
epoch: 9, loss_train: 0.5543, accuracy_train: 0.7920, loss_dev: 0.7401, accuracy_dev: 0.7026
epoch: 10, loss_train: 0.5499, accuracy_train: 0.7921, loss_dev: 0.7409, accuracy_dev: 0.7037