Files
paddlepaddle--paddle/test/auto_parallel/pir/test_reshard.py
T
2026-07-13 12:40:42 +08:00

103 lines
3.6 KiB
Python

# Copyright (c) 2024 PaddlePaddle Authors. All Rights Reserved.
#
# 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 pathlib
import sys
import unittest
import paddle
import paddle.distributed as dist
from paddle import nn
from paddle.distributed.auto_parallel.static.mix_to_dist_pass import (
apply_mix2dist_pass,
)
sys.path.append(str(pathlib.Path(__file__).resolve().parents[0]))
from test_to_static_pir_program import (
BATCH_SIZE,
CLASS_NUM,
DemoNet,
create_data_loader,
)
class ReshardDemoNet(DemoNet):
def __init__(self, mesh, shard=True):
super().__init__(mesh, shard=True)
def forward(self, x):
out = DemoNet.forward(self, x)
out = dist.reshard(out, self._mesh, [dist.Shard(0)])
return out
class TestToStaticPirProgramTrain(unittest.TestCase):
def test_to_static_program(self):
mesh = dist.ProcessMesh([0, 1], dim_names=["x"])
layer = ReshardDemoNet(mesh)
opt = paddle.optimizer.SGD(
learning_rate=0.1, parameters=layer.parameters()
)
loss_fn = nn.MSELoss()
loader = create_data_loader()
dist_loader = dist.shard_dataloader(loader, meshes=[mesh])
dist_model = dist.to_static(layer, dist_loader, loss_fn, opt)
engine = dist_model._engine
engine._build("train")
dist_program = engine._fwd_main_progs["train"]
apply_mix2dist_pass(dist_program)
loss = dist_program.get_output_value_by_name(engine._loss_names[0])
with paddle.static.program_guard(dist_program):
params_grads = paddle.autograd.ir_backward.append_backward(loss)
engine._optimizer._apply_optimize(
loss, startup_program=None, params_grads=params_grads
)
index = 0
for op in dist_program.global_block().ops:
if op.name() == 'dist_op.reshard':
if index == 0:
# forward reshard op
self.fwd_input = op.operand_source(0)
self.assertEqual(
self.fwd_input.dist_attr().dims_mapping, [-1, -1]
)
self.assertEqual(
self.fwd_input.dist_attr().partial_dims, set()
)
self.assertEqual(
self.fwd_input._local_shape,
[BATCH_SIZE, CLASS_NUM],
)
self.fwd_output = op.result(0)
self.assertEqual(
self.fwd_output.dist_attr().dims_mapping, [0, -1]
)
self.assertEqual(
self.fwd_output.dist_attr().partial_dims, set()
)
self.assertEqual(
self.fwd_output._local_shape,
[BATCH_SIZE / 2, CLASS_NUM],
)
elif index == 1:
# backward reshard op
self.assertEqual(op.result(0).type(), self.fwd_input.type())
index += 1
self.assertEqual(index, 2)
if __name__ == "__main__":
unittest.main()