wehub-resource-sync
71 lines
1.9 KiB
Python
71 lines
1.9 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 unittest
|
|
|
|
import paddle
|
|
from paddle.distributed.auto_parallel.intermediate.parallel_base import (
|
|
ParallelModel,
|
|
)
|
|
|
|
|
|
class PP(ParallelModel):
|
|
def __init__(self, model):
|
|
super().__init__(model)
|
|
self.pp_parallelizer = self.pp_init
|
|
|
|
def pp_init(self, model):
|
|
return paddle.nn.Linear(2, 2)
|
|
|
|
|
|
class TP(ParallelModel):
|
|
def __init__(self, model):
|
|
super().__init__(model)
|
|
self.tp_parallelizer = self.tp_init
|
|
|
|
def tp_init(self, model):
|
|
return paddle.nn.Linear(3, 3)
|
|
|
|
|
|
class SD(ParallelModel):
|
|
def __init__(self, model):
|
|
super().__init__(model)
|
|
self.sharding_parallelizer = self.sd_init
|
|
|
|
def sd_init(self, model):
|
|
return paddle.nn.Linear(4, 4)
|
|
|
|
|
|
class TestStrategy:
|
|
def test_recursive(self):
|
|
model = paddle.nn.Linear(1, 1)
|
|
pp = PP(model)
|
|
data = paddle.rand([1, 2])
|
|
pp(data)
|
|
assert pp.model.weight.shape == [2, 2]
|
|
model = paddle.nn.Linear(1, 1)
|
|
tp = TP(PP(model))
|
|
data = paddle.rand([1, 3])
|
|
tp(data)
|
|
assert tp.model.weight.shape == [3, 3]
|
|
model = paddle.nn.Linear(1, 1)
|
|
sd = SD(TP(PP(model)))
|
|
data = paddle.rand([1, 4])
|
|
sd(data)
|
|
assert sd.model.weight.shape == [4, 4]
|
|
|
|
|
|
if __name__ == '__main__':
|
|
unittest.main()
|