wehub-resource-sync
118 lines
3.6 KiB
Python
118 lines
3.6 KiB
Python
# Copyright (c) 2023 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 os
|
|
import unittest
|
|
|
|
import numpy as np
|
|
|
|
import paddle
|
|
import paddle.distributed as dist
|
|
|
|
|
|
class TestDistTensorLocalAPI(unittest.TestCase):
|
|
def setUp(self):
|
|
self._shape = eval(os.getenv("shape"))
|
|
self._dtype = os.getenv("dtype")
|
|
self._seed = 2023
|
|
self._backend = os.getenv("backend")
|
|
self._mesh = dist.ProcessMesh([0, 1], dim_names=["x"])
|
|
paddle.seed(self._seed)
|
|
np.random.seed(self._seed)
|
|
|
|
def run_test_dist_tensor_with_local_tensor_shard(self):
|
|
if self._backend == "cpu":
|
|
paddle.set_device("cpu")
|
|
place = paddle.CPUPlace()
|
|
elif self._backend == "gpu":
|
|
place = paddle.CUDAPlace(dist.get_rank())
|
|
|
|
global_tensor0 = paddle.rand([4, 10])
|
|
local_tensor_list0 = paddle.split(
|
|
global_tensor0, num_or_sections=2, axis=0
|
|
)
|
|
local_tensor0 = local_tensor_list0[dist.get_rank()]
|
|
|
|
dist_tensor_shard0 = dist.auto_parallel.api.dtensor_from_local(
|
|
local_tensor0,
|
|
mesh=self._mesh,
|
|
placements=[dist.Shard(0)],
|
|
)
|
|
|
|
np.testing.assert_equal(
|
|
dist_tensor_shard0._local_value().numpy(),
|
|
local_tensor0.numpy(),
|
|
)
|
|
np.testing.assert_equal(
|
|
dist_tensor_shard0.numpy(),
|
|
global_tensor0.numpy(),
|
|
)
|
|
|
|
self.assertEqual(dist_tensor_shard0.shape, [4, 10])
|
|
|
|
global_tensor1 = paddle.rand([2, 20])
|
|
local_tensor_list1 = paddle.split(
|
|
global_tensor1, num_or_sections=2, axis=1
|
|
)
|
|
local_tensor1 = local_tensor_list1[dist.get_rank()]
|
|
dist_tensor_shard1 = dist.auto_parallel.api.dtensor_from_local(
|
|
local_tensor1,
|
|
mesh=self._mesh,
|
|
placements=[dist.Shard(1)],
|
|
)
|
|
|
|
np.testing.assert_equal(
|
|
dist_tensor_shard1._local_value().numpy(),
|
|
local_tensor1.numpy(),
|
|
)
|
|
np.testing.assert_equal(
|
|
dist_tensor_shard1.numpy(),
|
|
global_tensor1.numpy(),
|
|
)
|
|
|
|
self.assertEqual(dist_tensor_shard1.shape, [2, 20])
|
|
|
|
def run_test_dist_tensor_with_local_tensor_replicate(self):
|
|
if self._backend == "cpu":
|
|
paddle.set_device("cpu")
|
|
place = paddle.CPUPlace()
|
|
elif self._backend == "gpu":
|
|
place = paddle.CUDAPlace(dist.get_rank())
|
|
|
|
local_tensor = paddle.rand([2, 10])
|
|
dist_tensor = dist.auto_parallel.api.dtensor_from_local(
|
|
local_tensor,
|
|
mesh=self._mesh,
|
|
placements=[dist.Replicate()],
|
|
)
|
|
|
|
np.testing.assert_equal(
|
|
dist_tensor._local_value().numpy(),
|
|
local_tensor.numpy(),
|
|
)
|
|
np.testing.assert_equal(
|
|
dist_tensor.numpy(),
|
|
local_tensor.numpy(),
|
|
)
|
|
|
|
self.assertEqual(dist_tensor.shape, [2, 10])
|
|
|
|
def test_case(self):
|
|
self.run_test_dist_tensor_with_local_tensor_shard()
|
|
self.run_test_dist_tensor_with_local_tensor_replicate()
|
|
|
|
|
|
if __name__ == "__main__":
|
|
unittest.main()
|