def test_cpu_restore_training(tmpdir):
"""Verify continue training session on CPU."""
tutils.reset_seed()
hparams = tutils.get_hparams()
model = LightningTestModel(hparams)
# logger file to get meta
test_logger_version = 10
logger = tutils.get_test_tube_logger(tmpdir, False, version=test_logger_version)
trainer_options = dict(
max_epochs=8,
val_check_interval=0.50,
val_percent_check=0.2,
train_percent_check=0.2,
logger=logger,
checkpoint_callback=ModelCheckpoint(tmpdir, save_top_k=-1)
)
# fit model
trainer = Trainer(**trainer_options)
result = trainer.fit(model)
real_global_epoch = trainer.current_epoch
# traning complete
assert result == 1, 'amp + ddp model failed to complete'
# wipe-out trainer and model
# retrain with not much data... this simulates picking training back up after slurm
# we want to see if the weights come back correctly
new_logger = tutils.get_test_tube_logger(tmpdir, False, version=test_logger_version)
trainer_options = dict(
max_epochs=2,
val_check_interval=0.50,
val_percent_check=0.2,
train_percent_check=0.2,
logger=new_logger,
checkpoint_callback=ModelCheckpoint(tmpdir),
)
trainer = Trainer(**trainer_options)
model = LightningTestModel(hparams)
# set the epoch start hook so we can predict before the model does the full training
def assert_good_acc():
assert trainer.current_epoch == real_global_epoch
assert trainer.current_epoch >= 0
# if model and state loaded correctly, predictions will be good even though we
# haven't trained with the new loaded model
trainer.model.eval()
for dataloader in trainer.get_val_dataloaders():
tutils.run_prediction(dataloader, trainer.model)
model.on_train_start = assert_good_acc
# by calling fit again, we trigger training, loading weights from the cluster
# and our hook to predict using current model before any more weight updates
trainer.fit(model)
def test_running_test_without_val(tmpdir):
"""Verify `test()` works on a model with no `val_loader`."""
tutils.reset_seed()
class CurrentTestModel(LightningTestMixin, LightningTestModelBase):
pass
hparams = tutils.get_hparams()
model = CurrentTestModel(hparams)
# logger file to get meta
logger = tutils.get_test_tube_logger(tmpdir, False)
# logger file to get weights
checkpoint = tutils.init_checkpoint_callback(logger)
trainer_options = dict(show_progress_bar=False,
max_epochs=1,
train_percent_check=0.4,
val_percent_check=0.2,
test_percent_check=0.2,
checkpoint_callback=checkpoint,
logger=logger)
# fit model
trainer = Trainer(**trainer_options)
result = trainer.fit(model)
assert result == 1, 'training failed to complete'
trainer.test()
# test we have good test accuracy
tutils.assert_ok_model_acc(trainer)
def test_running_test_pretrained_model(tmpdir):
tutils.reset_seed()
"""Verify test() on pretrained model"""
hparams = tutils.get_hparams()
model = LightningTestModel(hparams)
# logger file to get meta
logger = tutils.get_test_tube_logger(tmpdir, False)
# logger file to get weights
checkpoint = tutils.init_checkpoint_callback(logger)
trainer_options = dict(show_progress_bar=False,
max_num_epochs=4,
train_percent_check=0.4,
val_percent_check=0.2,
checkpoint_callback=checkpoint,
logger=logger)
# fit model
trainer = Trainer(**trainer_options)
result = trainer.fit(model)
# correct result and ok accuracy
assert result == 1, 'training failed to complete'
pretrained_model = tutils.load_model(logger.experiment,
trainer.checkpoint_callback.filepath,
module_class=LightningTestModel)
new_trainer = Trainer(**trainer_options)
new_trainer.test(pretrained_model)
# test we have good test accuracy
tutils.assert_ok_test_acc(new_trainer)
def test_ddp_sampler_error(tmpdir):
"""
Make sure DDP + AMP work
:return:
"""
if not tutils.can_run_gpu_test():
return
tutils.reset_seed()
tutils.set_random_master_port()
hparams = tutils.get_hparams()
model = LightningTestModel(hparams, force_remove_distributed_sampler=True)
logger = tutils.get_test_tube_logger(tmpdir, True)
trainer = Trainer(
logger=logger,
show_progress_bar=False,
max_nb_epochs=1,
gpus=[0, 1],
distributed_backend='ddp',
use_amp=True
)
with pytest.warns(UserWarning):
trainer.get_dataloaders(model)
def test_early_stopping_cpu_model(tmpdir):
"""Test each of the trainer options."""
tutils.reset_seed()
stopping = EarlyStopping(monitor='val_loss', min_delta=0.1)
trainer_options = dict(
default_save_path=tmpdir,
min_epochs=2,
early_stop_callback=stopping,
gradient_clip_val=1.0,
overfit_pct=0.20,
track_grad_norm=2,
print_nan_grads=True,
show_progress_bar=True,
logger=tutils.get_test_tube_logger(tmpdir),
train_percent_check=0.1,
val_percent_check=0.1,
)
model, hparams = tutils.get_model()
tutils.run_model_test(trainer_options,
model,
on_gpu=False,
early_stop=True)
# test freeze on cpu
model.freeze()
model.unfreeze()
def test_testtube_pickle():
"""
Verify that pickling a trainer containing a test tube logger works
"""
tutils.reset_seed()
hparams = tutils.get_hparams()
model = LightningTestModel(hparams)
save_dir = tutils.init_save_dir()
logger = tutils.get_test_tube_logger(False)
logger.log_hyperparams(hparams)
logger.save()
trainer_options = dict(max_nb_epochs=1,
train_percent_check=0.01,
logger=logger)
trainer = Trainer(**trainer_options)
pkl_bytes = pickle.dumps(trainer)
trainer2 = pickle.loads(pkl_bytes)
trainer2.logger.log_metrics({"acc": 1.0})
tutils.clear_save_dir()
def test_running_test_after_fitting(tmpdir):
"""Verify test() on fitted model."""
tutils.reset_seed()
hparams = tutils.get_hparams()
model = LightningTestModel(hparams)
# logger file to get meta
logger = tutils.get_test_tube_logger(tmpdir, False)
# logger file to get weights
checkpoint = tutils.init_checkpoint_callback(logger)
trainer_options = dict(default_save_path=tmpdir,
show_progress_bar=False,
max_epochs=4,
train_percent_check=0.4,
val_percent_check=0.2,
test_percent_check=0.2,
checkpoint_callback=checkpoint,
logger=logger)
# fit model
trainer = Trainer(**trainer_options)
result = trainer.fit(model)
assert result == 1, 'training failed to complete'
trainer.test()
# test we have good test accuracy
tutils.assert_ok_model_acc(trainer)
def test_no_val_end_module(tmpdir):
"""Tests use case where trainer saves the model, and user loads it from tags independently."""
tutils.reset_seed()
class CurrentTestModel(LightningValidationStepMixin,
LightningTestModelBase):
pass
hparams = tutils.get_hparams()
model = CurrentTestModel(hparams)
# logger file to get meta
logger = tutils.get_test_tube_logger(tmpdir, False)
trainer_options = dict(max_epochs=1,
logger=logger,
checkpoint_callback=ModelCheckpoint(tmpdir))
# fit model
trainer = Trainer(**trainer_options)
result = trainer.fit(model)
# traning complete
assert result == 1, 'amp + ddp model failed to complete'
# save model
new_weights_path = os.path.join(tmpdir, 'save_test.ckpt')
trainer.save_checkpoint(new_weights_path)
# load new model
tags_path = tutils.get_data_path(logger, path_dir=tmpdir)
tags_path = os.path.join(tags_path, 'meta_tags.csv')
model_2 = LightningTestModel.load_from_metrics(
weights_path=new_weights_path, tags_csv=tags_path)
model_2.eval()
def test_model_saving_loading():
"""
Tests use case where trainer saves the model, and user loads it from tags independently
:return:
"""
tutils.reset_seed()
hparams = tutils.get_hparams()
model = LightningTestModel(hparams)
save_dir = tutils.init_save_dir()
# logger file to get meta
logger = tutils.get_test_tube_logger(False)
trainer_options = dict(
max_nb_epochs=1,
logger=logger,
checkpoint_callback=ModelCheckpoint(save_dir)
)
# fit model
trainer = Trainer(**trainer_options)
result = trainer.fit(model)
# traning complete
assert result == 1, 'amp + ddp model failed to complete'
# make a prediction
for dataloader in model.test_dataloader():
for batch in dataloader:
break
x, y = batch
x = x.view(x.size(0), -1)
# generate preds before saving model
model.eval()
pred_before_saving = model(x)
# save model
new_weights_path = os.path.join(save_dir, 'save_test.ckpt')
trainer.save_checkpoint(new_weights_path)
# load new model
tags_path = logger.experiment.get_data_path(logger.experiment.name, logger.experiment.version)
tags_path = os.path.join(tags_path, 'meta_tags.csv')
model_2 = LightningTestModel.load_from_metrics(weights_path=new_weights_path,
tags_csv=tags_path)
model_2.eval()
# make prediction
# assert that both predictions are the same
new_pred = model_2(x)
assert torch.all(torch.eq(pred_before_saving, new_pred)).item() == 1
tutils.clear_save_dir()
def test_running_test_pretrained_model_ddp():
"""Verify test() on pretrained model"""
if not tutils.can_run_gpu_test():
return
tutils.reset_seed()
tutils.set_random_master_port()
hparams = tutils.get_hparams()
model = LightningTestModel(hparams)
save_dir = tutils.init_save_dir()
# exp file to get meta
logger = tutils.get_test_tube_logger(False)
# exp file to get weights
checkpoint = tutils.init_checkpoint_callback(logger)
trainer_options = dict(
show_progress_bar=False,
max_nb_epochs=1,
train_percent_check=0.4,
val_percent_check=0.2,
checkpoint_callback=checkpoint,
logger=logger,
gpus=[0, 1],
distributed_backend='ddp'
)
# fit model
trainer = Trainer(**trainer_options)
result = trainer.fit(model)
exp = logger.experiment
logging.info(os.listdir(exp.get_data_path(exp.name, exp.version)))
# correct result and ok accuracy
assert result == 1, 'training failed to complete'
pretrained_model = tutils.load_model(logger.experiment,
trainer.checkpoint_callback.filepath,
module_class=LightningTestModel)
# run test set
new_trainer = Trainer(**trainer_options)
new_trainer.test(pretrained_model)
for dataloader in model.test_dataloader():
tutils.run_prediction(dataloader, pretrained_model)
tutils.clear_save_dir()
def test_cpu_model(tmpdir):
"""Make sure model trains on CPU."""
tutils.reset_seed()
trainer_options = dict(default_save_path=tmpdir,
show_progress_bar=False,
logger=tutils.get_test_tube_logger(tmpdir),
max_epochs=1,
train_percent_check=0.4,
val_percent_check=0.4)
model, hparams = tutils.get_model()
tutils.run_model_test(trainer_options, model, on_gpu=False)
def test_cpu_model_with_amp(tmpdir):
"""Make sure model trains on CPU."""
tutils.reset_seed()
trainer_options = dict(default_save_path=tmpdir,
show_progress_bar=False,
logger=tutils.get_test_tube_logger(tmpdir),
max_epochs=1,
train_percent_check=0.4,
val_percent_check=0.4,
use_amp=True)
model, hparams = tutils.get_model()
with pytest.raises((MisconfigurationException, ModuleNotFoundError)):
tutils.run_model_test(trainer_options, model, on_gpu=False)
def test_cpu_model():
"""
Make sure model trains on CPU
:return:
"""
tutils.reset_seed()
trainer_options = dict(show_progress_bar=False,
logger=tutils.get_test_tube_logger(),
max_nb_epochs=1,
train_percent_check=0.4,
val_percent_check=0.4)
model, hparams = tutils.get_model()
tutils.run_model_test(trainer_options, model, hparams, on_gpu=False)
def test_testtube_logger(tmpdir):
"""Verify that basic functionality of test tube logger works."""
tutils.reset_seed()
hparams = tutils.get_hparams()
model = LightningTestModel(hparams)
logger = tutils.get_test_tube_logger(tmpdir, False)
trainer_options = dict(max_num_epochs=1,
train_percent_check=0.01,
logger=logger)
trainer = Trainer(**trainer_options)
result = trainer.fit(model)
assert result == 1, "Training failed"
def test_running_test_pretrained_model_dp():
tutils.reset_seed()
"""Verify test() on pretrained model"""
if not tutils.can_run_gpu_test():
return
hparams = tutils.get_hparams()
model = LightningTestModel(hparams)
save_dir = tutils.init_save_dir()
# logger file to get meta
logger = tutils.get_test_tube_logger(False)
# logger file to get weights
checkpoint = tutils.init_checkpoint_callback(logger)
trainer_options = dict(
show_progress_bar=True,
max_nb_epochs=1,
train_percent_check=0.4,
val_percent_check=0.2,
checkpoint_callback=checkpoint,
logger=logger,
gpus=[0, 1],
distributed_backend='dp'
)
# fit model
trainer = Trainer(**trainer_options)
result = trainer.fit(model)
# correct result and ok accuracy
assert result == 1, 'training failed to complete'
pretrained_model = tutils.load_model(logger.experiment,
trainer.checkpoint_callback.filepath,
module_class=LightningTestModel)
new_trainer = Trainer(**trainer_options)
new_trainer.test(pretrained_model)
# test we have good test accuracy
tutils.assert_ok_test_acc(new_trainer)
tutils.clear_save_dir()
def test_loading_meta_tags(tmpdir):
tutils.reset_seed()
from argparse import Namespace
hparams = tutils.get_hparams()
# save tags
logger = tutils.get_test_tube_logger(tmpdir, False)
logger.log_hyperparams(Namespace(some_str='a_str', an_int=1, a_float=2.0))
logger.log_hyperparams(hparams)
logger.save()
# load tags
path_expt_dir = tutils.get_data_path(logger, path_dir=tmpdir)
tags_path = os.path.join(path_expt_dir, 'meta_tags.csv')
tags = load_hparams_from_tags_csv(tags_path)
assert tags.batch_size == 32 and tags.hidden_dim == 1000
def test_loading_meta_tags(tmpdir):
tutils.reset_seed()
from argparse import Namespace
hparams = tutils.get_hparams()
# save tags
logger = tutils.get_test_tube_logger(tmpdir, False)
logger.log_hyperparams(Namespace(some_str='a_str', an_int=1, a_float=2.0))
logger.log_hyperparams(hparams)
logger.save()
# load tags
tags_path = logger.experiment.get_data_path(
logger.experiment.name, logger.experiment.version) + '/meta_tags.csv'
tags = training_io.load_hparams_from_tags_csv(tags_path)
assert tags.batch_size == 32 and tags.hidden_dim == 1000
def test_all_features_cpu_model(tmpdir):
"""Test each of the trainer options."""
tutils.reset_seed()
trainer_options = dict(default_save_path=tmpdir,
gradient_clip_val=1.0,
overfit_pct=0.20,
track_grad_norm=2,
print_nan_grads=True,
show_progress_bar=False,
logger=tutils.get_test_tube_logger(tmpdir),
accumulate_grad_batches=2,
max_epochs=1,
train_percent_check=0.4,
val_percent_check=0.4)
model, hparams = tutils.get_model()
tutils.run_model_test(trainer_options, model, on_gpu=False)
def test_no_val_module():
"""
Tests use case where trainer saves the model, and user loads it from tags independently
:return:
"""
tutils.reset_seed()
hparams = tutils.get_hparams()
class CurrentTestModel(LightningTestModelBase):
pass
model = CurrentTestModel(hparams)
save_dir = tutils.init_save_dir()
# logger file to get meta
logger = tutils.get_test_tube_logger(False)
trainer_options = dict(max_nb_epochs=1,
logger=logger,
checkpoint_callback=ModelCheckpoint(save_dir))
# fit model
trainer = Trainer(**trainer_options)
result = trainer.fit(model)
# training complete
assert result == 1, 'amp + ddp model failed to complete'
# save model
new_weights_path = os.path.join(save_dir, 'save_test.ckpt')
trainer.save_checkpoint(new_weights_path)
# load new model
tags_path = logger.experiment.get_data_path(logger.experiment.name,
logger.experiment.version)
tags_path = os.path.join(tags_path, 'meta_tags.csv')
model_2 = LightningTestModel.load_from_metrics(
weights_path=new_weights_path, tags_csv=tags_path)
model_2.eval()
# make prediction
tutils.clear_save_dir()
def test_amp_gpu_ddp_slurm_managed(tmpdir):
"""Make sure DDP + AMP work."""
if not tutils.can_run_gpu_test():
return
tutils.reset_seed()
# simulate setting slurm flags
tutils.set_random_master_port()
os.environ['SLURM_LOCALID'] = str(0)
hparams = tutils.get_hparams()
model = LightningTestModel(hparams)
trainer_options = dict(show_progress_bar=True,
max_epochs=1,
gpus=[0],
distributed_backend='ddp',
use_amp=True)
# exp file to get meta
logger = tutils.get_test_tube_logger(tmpdir, False)
# exp file to get weights
checkpoint = tutils.init_checkpoint_callback(logger)
# add these to the trainer options
trainer_options['checkpoint_callback'] = checkpoint
trainer_options['logger'] = logger
# fit model
trainer = Trainer(**trainer_options)
trainer.is_slurm_managing_tasks = True
result = trainer.fit(model)
# correct result and ok accuracy
assert result == 1, 'amp + ddp model failed to complete'
# test root model address
assert trainer.resolve_root_node_address('abc') == 'abc'
assert trainer.resolve_root_node_address('abc[23]') == 'abc23'
assert trainer.resolve_root_node_address('abc[23-24]') == 'abc23'
assert trainer.resolve_root_node_address(
'abc[23-24, 45-40, 40]') == 'abc23'
def test_amp_gpu_ddp_slurm_managed():
"""
Make sure DDP + AMP work
:return:
"""
if not tutils.can_run_gpu_test():
return
tutils.reset_seed()
# simulate setting slurm flags
tutils.set_random_master_port()
os.environ['SLURM_LOCALID'] = str(0)
hparams = tutils.get_hparams()
model = LightningTestModel(hparams)
trainer_options = dict(show_progress_bar=True,
max_nb_epochs=1,
gpus=[0],
distributed_backend='ddp',
use_amp=True)
save_dir = tutils.init_save_dir()
# exp file to get meta
logger = tutils.get_test_tube_logger(False)
# exp file to get weights
checkpoint = tutils.init_checkpoint_callback(logger)
# add these to the trainer options
trainer_options['checkpoint_callback'] = checkpoint
trainer_options['logger'] = logger
# fit model
trainer = Trainer(**trainer_options)
trainer.is_slurm_managing_tasks = True
result = trainer.fit(model)
# correct result and ok accuracy
assert result == 1, 'amp + ddp model failed to complete'
# test root model address
assert trainer.resolve_root_node_address('abc') == 'abc'
assert trainer.resolve_root_node_address('abc[23]') == 'abc23'
assert trainer.resolve_root_node_address('abc[23-24]') == 'abc23'
assert trainer.resolve_root_node_address(
'abc[23-24, 45-40, 40]') == 'abc23'
# test model loading with a map_location
pretrained_model = tutils.load_model(logger.experiment,
trainer.checkpoint_callback.filepath)
# test model preds
for dataloader in trainer.get_test_dataloaders():
tutils.run_prediction(dataloader, pretrained_model)
if trainer.use_ddp:
# on hpc this would work fine... but need to hack it for the purpose of the test
trainer.model = pretrained_model
trainer.optimizers, trainer.lr_schedulers = pretrained_model.configure_optimizers(
)
# test HPC loading / saving
trainer.hpc_save(save_dir, logger)
trainer.hpc_load(save_dir, on_gpu=True)
# test freeze on gpu
model.freeze()
model.unfreeze()
tutils.clear_save_dir()
def test_cpu_slurm_save_load(tmpdir):
"""
Verify model save/load/checkpoint on CPU
:return:
"""
tutils.reset_seed()
hparams = tutils.get_hparams()
model = LightningTestModel(hparams)
save_dir = tmpdir
# logger file to get meta
logger = tutils.get_test_tube_logger(save_dir, False)
version = logger.version
trainer_options = dict(
max_nb_epochs=1,
logger=logger,
checkpoint_callback=ModelCheckpoint(save_dir)
)
# fit model
trainer = Trainer(**trainer_options)
result = trainer.fit(model)
real_global_step = trainer.global_step
# traning complete
assert result == 1, 'amp + ddp model failed to complete'
# predict with trained model before saving
# make a prediction
for dataloader in model.test_dataloader():
for batch in dataloader:
break
x, y = batch
x = x.view(x.size(0), -1)
model.eval()
pred_before_saving = model(x)
# test HPC saving
# simulate snapshot on slurm
saved_filepath = trainer.hpc_save(save_dir, logger)
assert os.path.exists(saved_filepath)
# new logger file to get meta
logger = tutils.get_test_tube_logger(save_dir, False, version=version)
trainer_options = dict(
max_nb_epochs=1,
logger=logger,
checkpoint_callback=ModelCheckpoint(save_dir),
)
trainer = Trainer(**trainer_options)
model = LightningTestModel(hparams)
# set the epoch start hook so we can predict before the model does the full training
def assert_pred_same():
assert trainer.global_step == real_global_step and trainer.global_step > 0
# predict with loaded model to make sure answers are the same
trainer.model.eval()
new_pred = trainer.model(x)
assert torch.all(torch.eq(pred_before_saving, new_pred)).item() == 1
model.on_epoch_start = assert_pred_same
# by calling fit again, we trigger training, loading weights from the cluster
# and our hook to predict using current model before any more weight updates
trainer.fit(model)
def test_dp_resume():
"""
Make sure DP continues training correctly
:return:
"""
if not tutils.can_run_gpu_test():
return
tutils.reset_seed()
hparams = tutils.get_hparams()
model = LightningTestModel(hparams)
trainer_options = dict(
show_progress_bar=True,
max_nb_epochs=2,
gpus=2,
distributed_backend='dp',
)
save_dir = tutils.init_save_dir()
# get logger
logger = tutils.get_test_tube_logger(debug=False)
# exp file to get weights
# logger file to get weights
checkpoint = tutils.init_checkpoint_callback(logger)
# add these to the trainer options
trainer_options['logger'] = logger
trainer_options['checkpoint_callback'] = checkpoint
# fit model
trainer = Trainer(**trainer_options)
trainer.is_slurm_managing_tasks = True
result = trainer.fit(model)
# track epoch before saving
real_global_epoch = trainer.current_epoch
# correct result and ok accuracy
assert result == 1, 'amp + dp model failed to complete'
# ---------------------------
# HPC LOAD/SAVE
# ---------------------------
# save
trainer.hpc_save(save_dir, logger)
# init new trainer
new_logger = tutils.get_test_tube_logger(version=logger.version)
trainer_options['logger'] = new_logger
trainer_options['checkpoint_callback'] = ModelCheckpoint(save_dir)
trainer_options['train_percent_check'] = 0.2
trainer_options['val_percent_check'] = 0.2
trainer_options['max_nb_epochs'] = 1
new_trainer = Trainer(**trainer_options)
# set the epoch start hook so we can predict before the model does the full training
def assert_good_acc():
assert new_trainer.current_epoch == real_global_epoch and new_trainer.current_epoch > 0
# if model and state loaded correctly, predictions will be good even though we
# haven't trained with the new loaded model
dp_model = new_trainer.model
dp_model.eval()
dataloader = trainer.get_train_dataloader()
tutils.run_prediction(dataloader, dp_model, dp=True)
# new model
model = LightningTestModel(hparams)
model.on_sanity_check_start = assert_good_acc
# fit new model which should load hpc weights
new_trainer.fit(model)
# test freeze on gpu
model.freeze()
model.unfreeze()
tutils.clear_save_dir()