class CometExperimentLogger(ExperimentLogger):
def __init__(self, exp_name, online=True, **kwargs):
super(CometExperimentLogger, self).__init__(exp_name, **kwargs)
if online:
self.comet = Experiment(project_name=exp_name, **kwargs)
else:
self.comet = OfflineExperiment(project_name=exp_name, **kwargs)
def log_metric(self, tag, value, step, **kwargs):
self.comet.log_metric(tag, value, step=step, **kwargs)
def log_image(self, tag, img, step, **kwargs):
self.comet.log_image(img, name=tag, step=step, **kwargs)
def log_plt(self, tag, plt, step, **kwargs):
self.comet.log_figure(figure=plt, figure_name=tag, step=step, **kwargs)
def log_text(self, tag, text, **kwargs):
self.comet.log_text(text, **kwargs)
def log_parameters(self, params, **kwargs):
self.comet.log_parameters(params, **kwargs)
def start_epoch(self, **kwargs):
super(CometExperimentLogger, self).start_epoch()
def end_epoch(self, **kwargs):
super(CometExperimentLogger, self).end_epoch()
self.comet.log_epoch_end(self.epoch, **kwargs)
def end_experiment(self):
self.comet.end()
# Compute train accuracy
_, predicted = torch.max(outputs.data, 1)
total += labels.size(0)
correct += float((predicted == labels.data).sum())
# Log accuracy to Comet.ml
experiment.log_metric("accuracy", correct / total, step=step)
step += 1
if (i + 1) % 100 == 0:
print('Epoch [%d/%d], Step [%d/%d], Loss: %.4f' %
(epoch + 1, hyper_params['num_epochs'], i + 1,
len(train_dataset) // hyper_params['batch_size'],
loss.data.item()))
experiment.log_epoch_end(epoch)
with experiment.test():
# Test the Model
correct = 0
total = 0
for images, labels in test_loader:
images = Variable(
images.view(-1, hyper_params['sequence_length'],
hyper_params['input_size']))
outputs = rnn(images)
_, predicted = torch.max(outputs.data, 1)
total += labels.size(0)
correct += float((predicted == labels).sum())
experiment.log_metric("accuracy", correct / total)
print('Test Accuracy of the model on the 10000 test images: %d %%' %
