for parameter in model.parameters():
parameter.requires_grad = False
if (use_cuda):
model.cuda()
model.eval()
summary(model, (1, 64, 64))
print("Loaded model...")
preds = []
labels = []
for local_batch, local_labels in tqdm(test_generator):
labels.extend(local_labels.numpy().tolist())
local_batch, local_labels = local_batch.to(device), local_labels.to(device)
output = model.forward(local_batch)
output = model.softmax(output)
output = torch.max(output, 1)[1]
preds.extend(output.cpu().numpy().tolist())
recall = recall_score(y_true=labels, y_pred=preds, average='weighted')
prec = precision_score(y_true=labels, y_pred=preds, average='weighted')
f1 = f1_score(y_true=labels, y_pred=preds, average='weighted')
acc = accuracy_score(labels, preds)
print("Accuracy: {}".format(acc))
print("Recall: {}\tPrecision: {}\tF1 Score: {}".format(recall, prec, f1))
print(
classification_report(y_true=labels,
y_pred=preds,
target_names=["No Findings", "Pneumonia"]))
global_train_step = 0
global_val_step = 0
# Loop over epochs
for epoch in range(max_epochs):
tqdm.write("Epoch: {}".format(epoch))
progress_bar = tqdm(total=len(train_dataset), leave=True, position=0)
# Training
for local_batch, local_labels in training_generator:
# Transfer to GPU
local_batch, local_labels = local_batch.to(device), local_labels.to(
device)
loss, yhat = model.step(local_batch, local_labels)
yhat = model.softmax(yhat)
yhat = torch.max(yhat, 1)[1].cpu()
acc = pred_acc(local_labels.cpu(), yhat)
writer.add_scalar('Train/Accuracy', acc, global_train_step)
writer.add_scalar('Train/Loss', loss, global_train_step)
global_train_step += params['batch_size']
progress_bar.update(params['batch_size'])
progress_bar.set_postfix(loss=loss)
tqdm.write("Running validation...")
# Validation
y_pred = []
y_true = []
with torch.set_grad_enabled(False):
for local_batch, local_labels in validation_generator:
