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_amp_gpu_dp(tmpdir):
"""Make sure DP + AMP work."""
tutils.reset_seed()
if not tutils.can_run_gpu_test():
return
model, hparams = tutils.get_model()
trainer_options = dict(
default_save_path=tmpdir,
max_num_epochs=1,
gpus='0, 1', # test init with gpu string
distributed_backend='dp',
use_amp=True)
with pytest.raises(MisconfigurationException):
tutils.run_model_test(trainer_options, model, hparams)
def test_multi_gpu_none_backend(tmpdir):
"""Make sure when using multiple GPUs the user can't use `distributed_backend = None`."""
tutils.reset_seed()
if not tutils.can_run_gpu_test():
return
model, hparams = tutils.get_model()
trainer_options = dict(default_save_path=tmpdir,
show_progress_bar=False,
max_num_epochs=1,
train_percent_check=0.1,
val_percent_check=0.1,
gpus='-1')
with pytest.raises(MisconfigurationException):
tutils.run_model_test(trainer_options, model)
def test_multi_gpu_model_ddp(tmpdir):
"""Make sure DDP works."""
if not tutils.can_run_gpu_test():
return
tutils.reset_seed()
tutils.set_random_master_port()
model, hparams = tutils.get_model()
trainer_options = dict(default_save_path=tmpdir,
show_progress_bar=False,
max_num_epochs=1,
train_percent_check=0.4,
val_percent_check=0.2,
gpus=[0, 1],
distributed_backend='ddp')
tutils.run_model_test(trainer_options, model)
def test_amp_gpu_dp(tmpdir):
"""Make sure DP + AMP work."""
tutils.reset_seed()
if not tutils.can_run_gpu_test():
return
model, hparams = tutils.get_model()
trainer_options = dict(
default_save_path=tmpdir,
max_epochs=1,
gpus='0, 1', # test init with gpu string
distributed_backend='dp',
use_amp=True)
trainer = Trainer(**trainer_options)
result = trainer.fit(model)
assert result == 1
