def main():
# Load net
cnn = CNN()
loss_func = nn.MultiLabelSoftMarginLoss()
optimizer = optim.Adam(cnn.parameters(), lr=learning_rate)
if torch.cuda.is_available():
cnn.cuda()
loss_func.cuda()
# Load data
train_dataloader = dataset.get_train_data_loader()
test_dataloader = dataset.get_test_data_loader()
# Train model
for epoch in range(num_epochs):
cnn.train()
for i, (images, labels) in enumerate(train_dataloader):
images = Variable(images)
labels = Variable(labels.long())
if torch.cuda.is_available():
images = images.cuda()
labels = labels.cuda()
predict_labels = cnn(images)
loss = loss_func(predict_labels, labels)
optimizer.zero_grad()
loss.backward()
optimizer.step()
if (i + 1) % 100 == 0:
print("epoch:", epoch, "step:", i, "loss:", loss.item())
# Save and test model
if (epoch + 1) % 10 == 0:
filename = "model" + str(epoch + 1) + ".pkl"
torch.save(cnn.state_dict(), filename)
cnn.eval()
correct = 0
total = 0
for (image, label) in test_dataloader:
vimage = Variable(image)
if torch.cuda.is_available():
vimage = vimage.cuda()
output = cnn(vimage)
predict_label = ""
for k in range(4):
predict_label += config.CHAR_SET[np.argmax(
output[0, k * config.CHAR_SET_LEN:(k + 1) *
config.CHAR_SET_LEN].data.cpu().numpy())]
true_label = one_hot.vec2text(label.numpy()[0])
total += label.size(0)
if predict_label == true_label:
correct += 1
if total % 200 == 0:
print(
'Test Accuracy of the model on the %d test images: %f %%'
% (total, 100 * correct / total))
print('Test Accuracy of the model on the %d test images: %f %%' %
(total, 100 * correct / total))
print("save and test model...")
torch.save(cnn.state_dict(), "./model.pkl") # current is model.pkl
print("save last model")
def main():
args = parse_args()
twitter_csv_path = args.tweet_csv_file
device_type = args.device
use_bert = False
shuffle = False
train_data, dev_data, test_data = load_twitter_data(twitter_csv_path, test_split_percent=0.1, val_split_percent=0.2, overfit=True, shuffle=shuffle, use_bert=use_bert, overfit_val=12639)
vocab_size = train_data.vocab_size
print(vocab_size)
print(train_data.length)
print(dev_data.length)
print(test_data.length)
cnn_net = CNN(vocab_size, DIM_EMB=300, NUM_CLASSES = 2)
if device_type == "gpu" and torch.cuda.is_available():
device = torch.device('cuda:0')
cnn_net = cnn_net.cuda()
epoch_losses, eval_accuracy = train_network(cnn_net,
train_data.Xwordlist,
(train_data.labels + 1.0)/2.0,
10, dev_data, lr=0.003,
batchSize=150, use_gpu=True, device=device)
cnn_net.eval()
print("Test Set")
test_accuracy = eval_network(test_data, cnn_net, use_gpu=True, device=device)
else:
device = torch.device('cpu')
epoch_losses, eval_accuracy = train_network(cnn_net,
train_data.Xwordlist,
(train_data.labels + 1.0)/2.0,
10, dev_data, lr=0.003,
batchSize=150, use_gpu=False, device=device)
cnn_net.eval()
print("Test Set")
test_accuracy = eval_network(test_data, cnn_net, use_gpu=False, batch_size=batchSize, device=device)
# plot_accuracy((min_accs, eval_accuracy, max_accs), "Sentiment CNN lr=0.001", train_data.length)
plot_accuracy(eval_accuracy, "Sentiment CNN lr=0.003", train_data.length)
plot_losses(epoch_losses, "Sentiment CNN lr=0.003", train_data.length)
torch.save(cnn_net.state_dict(), "saved_models\\cnn.pth")
np.save("cnn_train_loss_" + str(train_data.length) + ".npy", np.array(epoch_losses))
np.save("cnn_validation_accuracy_" + str(train_data.length) + ".npy", np.array(eval_accuracy))
def main():
args = parse_args()
# twitter_csv_path = args.tweet_csv_file
labeled_twitter_csv_path = args.labeled_tweet_csv_file
unlabeled_twitter_csv_path = args.unlabeled_tweet_csv_file
device_type = args.device
acquistion_function_type = args.acquisition_func
human_label = args.human_label
use_model_acq = True #flag for using model to generate inputs for acquisition funciton
if acquistion_function_type == "least_confidence":
acquisition_func = least_confidence
elif acquistion_function_type == "random":
acquisition_func = random_score
elif acquistion_function_type == "entropy":
acquisition_func = entropy_score
elif acquistion_function_type == "tweet_count":
acquisition_func = tweet_count_norm
use_model_acq = False
else:
acquisition_func = least_confidence
seed_data_size = args.seed_data_size
use_bert = False
shuffle = False
train_data, dev_data, test_data = load_twitter_data(
labeled_twitter_csv_path,
test_split_percent=0.1,
val_split_percent=0.2,
shuffle=shuffle,
overfit=True,
use_bert=use_bert,
overfit_val=40000)
unlabeled_tweets, ground_truth_labels = load_unlabeled_tweet_csv(
unlabeled_twitter_csv_path, num_tweets=45000)
#convert "unlabeled" tweets to token ids
X_unlabeled = train_data.convert_text_to_ids(unlabeled_tweets)
# ground_truth_labels = ground_truth_labels[0:70000]
ground_truth_labels = (ground_truth_labels + 1.0) / 2.0
X_seed = train_data.Xwordlist[0:seed_data_size]
Y_seed = train_data.labels[0:seed_data_size]
Y_seed = (Y_seed + 1.0) / 2.0
print(train_data.vocab_size)
print(len(X_seed))
print(dev_data.length)
print(test_data.length)
num_samples = args.sample_size
cnn_net = CNN(train_data.vocab_size, DIM_EMB=300, NUM_CLASSES=2)
if device_type == "gpu" and torch.cuda.is_available():
device = torch.device('cuda:0')
cnn_net = cnn_net.cuda()
epoch_losses, eval_accuracy, hand_labeled_data = train_active_learning(
cnn_net,
train_data,
X_seed,
Y_seed,
X_unlabeled,
ground_truth_labels,
dev_data,
use_model=use_model_acq,
num_epochs=8,
human_label=human_label,
acquisition_func=acquisition_func,
lr=0.0035,
batchSize=150,
num_samples=num_samples,
use_gpu=True,
device=device)
cnn_net.eval()
print("Test Set")
test_accuracy = eval_network(test_data,
cnn_net,
use_gpu=True,
device=device)
else:
device = torch.device('cpu')
# cnn_net = cnn_net.cuda()
epoch_losses, eval_accuracy, hand_labeled_data = train_active_learning(
cnn_net,
train_data,
X_seed,
Y_seed,
X_unlabeled,
ground_truth_labels,
dev_data,
use_model=use_model_acq,
num_epochs=8,
human_label=human_label,
acquisition_func=acquisition_func,
lr=0.0035,
batchSize=150,
num_samples=num_samples,
use_gpu=False,
device=device)
cnn_net.eval()
print("Test Set")
test_accuracy = eval_network(test_data,
cnn_net,
use_gpu=False,
device=device)
# plot_accuracy((min_accs, eval_accuracy, max_accs), "Sentiment CNN lr=0.001", train_data.length)
plot_accuracy(
eval_accuracy, "Sentiment CNN (Active Learning) lr=0.0035 " +
acquistion_function_type, seed_data_size)
# plot_losses(epoch_losses, "Sentiment CNN (Active Learning) lr=0.0030" + acquistion_function_type, train_data.length)
torch.save(cnn_net.state_dict(), "saved_models\\cnn_active_learn.pth")
# np.save("cnn_active_learning_train_loss" + acquistion_function_type + "_" + str(seed_data_size) + ".npy", np.array(epoch_losses))
np.save(
"human_labelling_results/cnn_active_learning_validation_accuracy_" +
acquistion_function_type + "_" + str(seed_data_size) + "_" +
str(num_samples) + ".npy", np.array(eval_accuracy))
human_labels = []
ground_truth_labels = []
tweets = []
save_labels = True
if save_labels:
for tweet, label, ground_truth_label in hand_labeled_data:
# tweet, score = sample
tweet = train_data.convert_to_words(tweet)
tweets.append(tweet)
human_labels.append(label)
ground_truth_labels.append(ground_truth_label)
new_labeled_tweets = pd.DataFrame({
'label': human_labels,
'ground truth': ground_truth_labels,
'text': tweets
})
new_labeled_tweets.to_csv("human_labeled_tweets_lc_rk.csv",
header=True,
index=False)
def main():
#parameters
# sampling_functions = ['random_score', 'entropy_score', 'least_confidence']
sampling_functions = ['tweet_count']
sampling_sizes = [5000, 10000, 15000, 20000]
num_active_samples = [10, 25, 50]
# sampling_functions = ['least_confidence']
# num_active_samples = [25, 50]
# sampling_sizes = [20000]
args = parse_args()
# twitter_csv_path = args.tweet_csv_file
labeled_twitter_csv_path = args.labeled_tweet_csv_file
unlabeled_twitter_csv_path = args.unlabeled_tweet_csv_file
save_models = args.save_models
use_bert = False
shuffle = False
train_data, dev_data, test_data = load_twitter_data(labeled_twitter_csv_path,
test_split_percent=0.1,
val_split_percent=0.2,
shuffle=shuffle,
overfit=True, use_bert=use_bert,
overfit_val=40000)
unlabeled_tweets, ground_truth_labels = load_unlabeled_tweet_csv(unlabeled_twitter_csv_path, num_tweets=45000)
X_unlabeled = train_data.convert_text_to_ids(unlabeled_tweets)
ground_truth_labels = ground_truth_labels
ground_truth_labels = (ground_truth_labels + 1.0)/2.0
test_accuracies = {}
print("Running ablation experiment on sampling functions and seed sizes")
use_model=True
for af in sampling_functions:
if af == 'random_score':
acquisition_func = random_score
elif af == 'entropy_score':
acquisition_func = entropy_score
elif af == 'least_confidence':
acquisition_func = least_confidence
elif af == 'tweet_count':
acquisition_func = tweet_count_norm
use_model=False
for seed_data_size in sampling_sizes:
for sample_size in num_active_samples:
param_combo = "Acquisition_Func: " + af + " Seed Size: " + str(seed_data_size) + " Sample Size: " + str(sample_size)
print(param_combo + "\n")
X_seed = train_data.Xwordlist[0:seed_data_size]
Y_seed = train_data.labels[0:seed_data_size]
Y_seed = (Y_seed + 1.0)/2.0
cnn_net = CNN(train_data.vocab_size, DIM_EMB=300, NUM_CLASSES = 2)
device = torch.device('cuda:0')
cnn_net = cnn_net.cuda()
print("Train active learning")
epoch_losses, eval_accuracy, hand_labeled_data = train_active_learning(cnn_net, train_data,
X_seed, Y_seed,
copy.deepcopy(X_unlabeled), np.copy(ground_truth_labels), dev_data,
num_epochs=8, use_model=use_model, acquisition_func=acquisition_func,
lr=0.0035, batchSize=150, num_samples=sample_size,
use_gpu=True, device=device)
print("Finished Training")
cnn_net.eval()
print("Test Set")
test_accuracy = eval_network(test_data, cnn_net, use_gpu=True, device=device)
model_save_path = "model_weights/cnn_active_learn_weights_"+ af + "_" + str(seed_data_size) + "_" + str(sample_size) + ".pth"
if save_models:
torch.save(cnn_net.state_dict(), model_save_path)
param_combo = "CNN Active Learning: " + " Acquisition_Func: " + af + " Seed Size: " + str(seed_data_size) + " Sample Size: " + str(sample_size)
test_accuracies[param_combo] = test_accuracy
filename = "results_ablation/cnn_active_learning_val_accuracy_" + af + "_" + str(seed_data_size) + "_" + str(sample_size) + ".npy"
np.save(filename, np.array(eval_accuracy))
print("Finished experiments")
with open("ablation_test_accuracies1.txt", "w") as f:
for key in test_accuracies.keys():
accuracy = test_accuracies[key]
line = key + " Acc: " + str(accuracy) + "\n"
f.write(line)
def train():
"""
Performs training and evaluation of MLP cnn.
NOTE: You should the cnn on the whole test set each eval_freq iterations.
"""
# YOUR TRAINING CODE GOES HERE
transform = transforms.Compose([
transforms.ToTensor(),
transforms.Normalize((0.5, 0.5, 0.5), (0.5, 0.5, 0.5))
])
train_data = datasets.CIFAR10('data',
train=True,
download=True,
transform=transform)
test_data = datasets.CIFAR10('data',
train=False,
download=True,
transform=transform)
train_on_gpu = torch.cuda.is_available()
num_train = len(train_data)
train_loader = torch.utils.data.DataLoader(train_data,
batch_size=FLAGS.batch_size,
shuffle=True,
num_workers=0)
test_loader = torch.utils.data.DataLoader(test_data,
batch_size=FLAGS.batch_size,
shuffle=False,
num_workers=0)
classes = [
'airplane', 'automobile', 'bird', 'cat', 'deer', 'dog', 'frog',
'horse', 'ship', 'truck'
]
cnn = CNN(3, 10)
if train_on_gpu:
cnn.cuda()
criterion = nn.CrossEntropyLoss()
optimizer = optim.Adam(cnn.parameters(), lr=FLAGS.learning_rate)
for epoch in range(1, FLAGS.max_steps):
class_correct = list(0. for i in range(10))
class_total = list(0. for i in range(10))
train_loss = 0.0
test_loss = 0.0
cnn.train()
for data, target in train_loader:
if train_on_gpu:
data, target = data.cuda(), target.cuda()
optimizer.zero_grad()
output = cnn(data)
loss = criterion(output, target)
loss.backward()
optimizer.step()
_, pred = torch.max(output, 1)
correct_tensor = pred.eq(target.data.view_as(pred))
train_loss += loss.item() * data.size(0)
correct = np.squeeze(
correct_tensor.numpy()) if not train_on_gpu else np.squeeze(
correct_tensor.cpu().numpy())
for i in range(len(target.data)):
label = target.data[i]
class_correct[label] += correct[i].item()
class_total[label] += 1
if epoch % FLAGS.eval_freq == 0:
test_correct = list(0. for i in range(10))
test_total = list(0. for i in range(10))
cnn.eval()
for data, target in test_loader:
if train_on_gpu:
data, target = data.cuda(), target.cuda()
output = cnn(data)
_, pred = torch.max(output, 1)
correct_tensor = pred.eq(target.data.view_as(pred))
correct = np.squeeze(correct_tensor.numpy()
) if not train_on_gpu else np.squeeze(
correct_tensor.cpu().numpy())
loss = criterion(output, target)
test_loss += loss.item() * data.size(0)
for i in range(len(target.data)):
label = target.data[i]
test_correct[label] += correct[i].item()
test_total[label] += 1
train_loss = train_loss / len(train_loader.dataset)
test_loss = test_loss / len(test_loader.dataset)
plot_epoch.append(epoch)
plot_train_loss.append(train_loss)
plot_test_loss.append(test_loss)
print(
'Epoch: {} \tTraining Loss: {:.6f} \tTest Loss: {:.6f}'.format(
epoch, train_loss, test_loss))
percent_train = accuracy(class_correct, class_total) * 100
percent_test = accuracy(test_correct, test_total) * 100
plot_train_accuracy.append(percent_train)
plot_test_accuracy.append(percent_test)
print('train accuracy: ', percent_train, 'test accuracy: ',
percent_test)
fig1 = plt.subplot(2, 1, 1)
fig2 = plt.subplot(2, 1, 2)
fig1.plot(plot_epoch,
plot_train_accuracy,
c='red',
label='training data accuracy')
fig1.plot(plot_epoch,
plot_test_accuracy,
c='blue',
label='test data accuracy')
fig1.legend()
fig2.plot(plot_epoch, plot_train_loss, c='green', label='train CE loss')
fig2.plot(plot_epoch, plot_test_loss, c='yellow', label='test CE loss')
fig2.legend()
plt.show()