72 lines
2.1 KiB
Python
72 lines
2.1 KiB
Python
|
import pytest
|
||
|
|
import torch
|
||
|
|
|
||
|
|
import tests.base.develop_pipelines as tpipes
|
||
|
|
import tests.base.develop_utils as tutils
|
||
|
|
from tests.base import EvalModelTemplate
|
||
|
|
from pytorch_lightning.core import memory
|
||
|
|
from pytorch_lightning.trainer import Trainer
|
||
|
|
|
||
|
|
|
||
|
|
@pytest.mark.skipif(torch.cuda.device_count() < 2, reason="test requires multi-GPU machine")
|
||
|
|
def test_multi_gpu_early_stop_ddp_spawn(tmpdir):
|
||
|
|
"""Make sure DDP works. with early stopping"""
|
||
|
|
tutils.set_random_master_port()
|
||
|
|
|
||
|
|
trainer_options = dict(
|
||
|
|
default_root_dir=tmpdir,
|
||
|
|
early_stop_callback=True,
|
||
|
|
max_epochs=50,
|
||
|
|
limit_train_batches=10,
|
||
|
|
limit_val_batches=10,
|
||
|
|
gpus=[0, 1],
|
||
|
|
distributed_backend='ddp_spawn',
|
||
|
|
)
|
||
|
|
|
||
|
|
model = EvalModelTemplate()
|
||
|
|
tpipes.run_model_test(trainer_options, model)
|
||
|
|
|
||
|
|
|
||
|
|
@pytest.mark.skipif(torch.cuda.device_count() < 2, reason="test requires multi-GPU machine")
|
||
|
|
def test_multi_gpu_model_ddp_spawn(tmpdir):
|
||
|
|
tutils.set_random_master_port()
|
||
|
|
|
||
|
|
trainer_options = dict(
|
||
|
|
default_root_dir=tmpdir,
|
||
|
|
max_epochs=1,
|
||
|
|
limit_train_batches=10,
|
||
|
|
limit_val_batches=10,
|
||
|
|
gpus=[0, 1],
|
||
|
|
distributed_backend='ddp_spawn',
|
||
|
|
progress_bar_refresh_rate=0
|
||
|
|
)
|
||
|
|
|
||
|
|
model = EvalModelTemplate()
|
||
|
|
|
||
|
|
tpipes.run_model_test(trainer_options, model)
|
||
|
|
|
||
|
|
# test memory helper functions
|
||
|
|
memory.get_memory_profile('min_max')
|
||
|
|
|
||
|
|
|
||
|
|
@pytest.mark.skipif(torch.cuda.device_count() < 2, reason="test requires multi-GPU machine")
|
||
|
|
def test_ddp_all_dataloaders_passed_to_fit(tmpdir):
|
||
|
|
"""Make sure DDP works with dataloaders passed to fit()"""
|
||
|
|
tutils.set_random_master_port()
|
||
|
|
|
||
|
|
model = EvalModelTemplate()
|
||
|
|
fit_options = dict(train_dataloader=model.train_dataloader(),
|
||
|
|
val_dataloaders=model.val_dataloader())
|
||
|
|
|
||
|
|
trainer = Trainer(
|
||
|
|
default_root_dir=tmpdir,
|
||
|
|
progress_bar_refresh_rate=0,
|
||
|
|
max_epochs=1,
|
||
|
|
limit_train_batches=0.2,
|
||
|
|
limit_val_batches=0.2,
|
||
|
|
gpus=[0, 1],
|
||
|
|
distributed_backend='ddp_spawn'
|
||
|
|
)
|
||
|
|
result = trainer.fit(model, **fit_options)
|
||
|
|
assert result == 1, "DDP doesn't work with dataloaders passed to fit()."
|