def test_one_batch_update(example_batch):
data, loss, output, target, batch_size = \
example_batch.data, example_batch.loss, \
example_batch.output, example_batch.target, example_batch.batch_size
arg_list = ['--no-save', '--no-visualize', '--epochs=1']
args, _, _ = init_pipeline(arg_list)
metrics = MetricTracker(args, {})
tqdm_dict = metrics.batch_update(0, 1, batch_size, data, loss, output, target, Mode.TRAIN)
for key in tqdm_dict:
tqdm_dict[key] = round(tqdm_dict[key], 2)
result = [round(metric.epoch_avg, 2) for metric in metrics.metric_data.values()]
assert tqdm_dict == {'Loss': 0.21, 'Accuracy': 0.67}
assert result == [0.63, 2.0]
def test_init_metrics():
arg_list = ["--no-save", "--no-visualize", "--num-examples=100", "--epochs=1"]
args, _, _ = init_pipeline(arg_list)
metrics = MetricTracker(args, {})
assert metrics
def test_epoch_update(capsys, example_batch):
data, loss, output, target, batch_size = \
example_batch.data, example_batch.loss, \
example_batch.output, example_batch.target, example_batch.batch_size
arg_list = ['--no-save', '--no-visualize', '--epochs=1', '--log-interval=3']
args, _, _ = init_pipeline(arg_list)
metrics = MetricTracker(args, {})
num_batches = 4
for i in range(num_batches):
_ = metrics.batch_update(i, num_batches, batch_size,
data, loss, output, target, Mode.TRAIN)
metrics.epoch_update(Mode.TRAIN)
captured = capsys.readouterr().out
assert captured == 'Mode.TRAIN Loss: 0.2100 Accuracy: 66.67% \n'
def init_metrics(args, checkpoint):
run_name = checkpoint.get('run_name', util.get_run_name(args))
metric_checkpoint = checkpoint.get('metric_obj', {})
metrics = MetricTracker(args.metric_names, run_name, args.log_interval,
**metric_checkpoint)
with open(os.path.join(run_name, 'args.json'),
'w') as f: # Save used args to checkpoint folder
json.dump(args.__dict__, f, indent=4)
return run_name, metrics
def test_many_batch_update(example_batch):
data, loss, output, target, batch_size = \
example_batch.data, example_batch.loss, \
example_batch.output, example_batch.target, example_batch.batch_size
arg_list = ['--no-save', '--no-visualize', '--epochs=1', '--log-interval=3']
args, _, _ = init_pipeline(arg_list)
metrics = MetricTracker(args, {})
num_batches = 4
for i in range(num_batches):
tqdm_dict = metrics.batch_update(i, num_batches, batch_size,
data, loss, output, target, Mode.TRAIN)
for key in tqdm_dict:
tqdm_dict[key] = round(tqdm_dict[key], 2)
result = [round(metric.epoch_avg, 2) for metric in metrics.metric_data.values()]
assert tqdm_dict == {'Loss': 0.21, 'Accuracy': 0.67}
assert all([metric.running_avg == 0. for metric in metrics.metric_data.values()])
assert result == [2.52, 8.]
def train(arg_list=None):
args, device, checkpoint = init_pipeline(arg_list)
train_loader, val_loader, init_params = load_train_data(args, device)
sample_loader = util.get_sample_loader(train_loader)
model, criterion, optimizer, scheduler = load_model(
args, device, init_params, sample_loader)
util.load_state_dict(checkpoint, model, optimizer, scheduler)
metrics = MetricTracker(args, checkpoint)
if not args.no_visualize:
metrics.add_network(model, sample_loader)
visualize(model, sample_loader, metrics.run_name)
util.set_rng_state(checkpoint)
for _ in range(args.epochs):
metrics.next_epoch()
train_and_validate(args, model, train_loader, optimizer, criterion,
metrics, Mode.TRAIN)
train_and_validate(args, model, val_loader, None, criterion, metrics,
Mode.VAL)
if args.scheduler:
scheduler.step()
if not args.no_save:
util.save_checkpoint(
{
"model_init":
init_params,
"model_state_dict":
model.state_dict(),
"optimizer_state_dict":
optimizer.state_dict(),
"scheduler_state_dict":
scheduler.state_dict() if args.scheduler else None,
"rng_state":
random.getstate(),
"np_rng_state":
np.random.get_state(),
"torch_rng_state":
torch.get_rng_state(),
"run_name":
metrics.run_name,
"metric_obj":
metrics.json_repr(),
},
metrics.is_best,
)
torch.set_grad_enabled(True)
if not args.no_visualize:
visualize_trained(model, sample_loader, metrics.run_name)
return metrics