lightning/tests/tests_fabric/strategies/test_ddp_integration.py

136 lines
5.3 KiB
Python

# Copyright The Lightning AI team.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
import os
import sys
from copy import deepcopy
from unittest import mock
from unittest.mock import Mock
import pytest
import torch
from lightning.fabric import Fabric
from lightning.fabric.utilities.imports import _TORCH_GREATER_EQUAL_2_0, _TORCH_GREATER_EQUAL_2_2
from torch.nn.parallel.distributed import DistributedDataParallel
from tests_fabric.helpers.runif import RunIf
from tests_fabric.strategies.test_single_device import _run_test_clip_gradients
from tests_fabric.test_fabric import BoringModel
@pytest.mark.xfail(
# https://github.com/pytorch/pytorch/issues/116056
sys.platform == "win32" and _TORCH_GREATER_EQUAL_2_2,
reason="Windows + DDP issue in PyTorch 2.2",
)
@pytest.mark.parametrize(
"accelerator",
[
"cpu",
pytest.param("cuda", marks=RunIf(min_cuda_gpus=2)),
],
)
def test_ddp_save_load(accelerator, tmp_path):
"""Test that DDP model checkpoints can be saved and loaded successfully."""
fabric = Fabric(devices=2, accelerator=accelerator, strategy="ddp_spawn")
fabric.launch(_run_ddp_save_load, tmp_path)
def _run_ddp_save_load(fabric, tmp_path):
fabric.seed_everything(0)
tmp_path = fabric.broadcast(tmp_path)
model = torch.nn.Linear(2, 2)
params_before = deepcopy(list(model.parameters()))
# Save
fabric.save(tmp_path / "saved_before_setup.ckpt", {"model": model})
wrapped_model = fabric.setup(model)
fabric.save(tmp_path / "saved_after_setup.ckpt", {"model": wrapped_model})
def assert_params_equal(params0, params1):
assert all(torch.equal(p0, p1.to(p0.device)) for p0, p1 in zip(params0, params1))
# Load
model = torch.nn.Linear(2, 2)
fabric.load(tmp_path / "saved_before_setup.ckpt", {"model": model})
assert_params_equal(params_before, model.parameters())
fabric.load(tmp_path / "saved_after_setup.ckpt", {"model": model})
assert_params_equal(params_before, model.parameters())
wrapped_model = fabric.setup(model)
fabric.load(tmp_path / "saved_before_setup.ckpt", {"model": wrapped_model})
assert_params_equal(params_before, wrapped_model.parameters())
fabric.load(tmp_path / "saved_after_setup.ckpt", {"model": wrapped_model})
assert_params_equal(params_before, wrapped_model.parameters())
@RunIf(min_cuda_gpus=2, standalone=True, min_torch="2.1.0", dynamo=True)
@mock.patch(
"lightning.fabric.wrappers.torch.compile",
Mock(wraps=(torch.compile if _TORCH_GREATER_EQUAL_2_0 else None)),
)
@mock.patch.dict(os.environ, {})
def test_reapply_compile():
"""Test that Fabric can rewrap a compiled module such that compilation happens over the DDP-wrapper."""
from torch._dynamo import OptimizedModule
fabric = Fabric(accelerator="cuda", devices=2, strategy="ddp")
fabric.launch()
model = BoringModel()
compile_kwargs = {"mode": "reduce-overhead"}
compiled_model = torch.compile(model, **compile_kwargs)
torch.compile.reset_mock()
fabric_model = fabric.setup(compiled_model, _reapply_compile=True)
assert isinstance(fabric_model._forward_module, OptimizedModule)
assert isinstance(fabric_model._forward_module._orig_mod, DistributedDataParallel)
# Assert we called compile again with the same arguments, but on the DDP-wrapped module
torch.compile.assert_called_with(fabric_model._forward_module._orig_mod, **compile_kwargs)
assert fabric_model._original_module == model
assert fabric_model._forward_module._orig_mod.module == model
assert fabric_model.device == fabric.device
# Smoke-testing forward to ensure we don't get compilation errors
for _ in range(3):
fabric_model(torch.randn(2, 32, device=fabric.device)).sum().backward()
@pytest.mark.parametrize(
("clip_type", "accelerator", "precision"),
[
("norm", "cpu", "32-true"),
("val", "cpu", "32-true"),
("norm", "cpu", "bf16-mixed"),
("val", "cpu", "bf16-mixed"),
pytest.param("norm", "cuda", "32-true", marks=RunIf(min_cuda_gpus=2)),
pytest.param("val", "cuda", "32-true", marks=RunIf(min_cuda_gpus=2)),
pytest.param("norm", "cuda", "16-mixed", marks=RunIf(min_cuda_gpus=2)),
pytest.param("val", "cuda", "16-mixed", marks=RunIf(min_cuda_gpus=2)),
pytest.param("norm", "cuda", "bf16-mixed", marks=RunIf(min_cuda_gpus=2, bf16_cuda=True)),
pytest.param("val", "cuda", "bf16-mixed", marks=RunIf(min_cuda_gpus=2, bf16_cuda=True)),
],
)
@RunIf(standalone=True)
def test_clip_gradients(clip_type, accelerator, precision):
if clip_type == "norm" and precision == "16-mixed":
pytest.skip(reason="Clipping by norm with 16-mixed is numerically unstable.")
fabric = Fabric(accelerator=accelerator, devices=2, precision=precision, strategy="ddp")
fabric.launch()
_run_test_clip_gradients(fabric=fabric, clip_type=clip_type)