152 lines
6.5 KiB
Python
152 lines
6.5 KiB
Python
# Copyright 2022 Huawei Technologies Co., Ltd
|
|
#
|
|
# 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 numpy as np
|
|
import pytest
|
|
import mindspore.context as context
|
|
import mindspore.nn as nn
|
|
from mindspore import Tensor
|
|
from mindspore import dtype as mstype
|
|
from mindspore.ops.operations.sparse_ops import Sspaddmm
|
|
|
|
|
|
class SspaddmmNet(nn.Cell):
|
|
|
|
def __init__(self):
|
|
super(SspaddmmNet, self).__init__()
|
|
self.sspaddmm = Sspaddmm()
|
|
|
|
def construct(self, x1_indices, x1_values, x1_shape, x2_indices, x2_values,
|
|
x2_shape, x3_dense, alpha, beta):
|
|
return self.sspaddmm(x1_indices, x1_values, x1_shape, x2_indices,
|
|
x2_values, x2_shape, x3_dense, alpha, beta)
|
|
|
|
|
|
@pytest.mark.levle0
|
|
@pytest.mark.platform_x86_gpu
|
|
@pytest.mark.env_onecard
|
|
def test_sspaddmm_dyn():
|
|
"""
|
|
Feature: test Sspaddmm ops in gpu.
|
|
Description: test the ops in dynamic shape.
|
|
Expectation: expect correct shape result.
|
|
"""
|
|
context.set_context(mode=context.GRAPH_MODE, device_target='GPU')
|
|
net = SspaddmmNet()
|
|
|
|
x1_indices_dyn = Tensor(shape=[2, None], dtype=mstype.int64)
|
|
x1_values_dyn = Tensor(shape=[None], dtype=mstype.int32)
|
|
x1_shape_dyn = Tensor(shape=[None], dtype=mstype.int64)
|
|
x2_indices_dyn = Tensor(shape=[None, None], dtype=mstype.int64)
|
|
x2_values_dyn = Tensor(shape=[None], dtype=mstype.int32)
|
|
x2_shape_dyn = Tensor(shape=[None], dtype=mstype.int64)
|
|
x3_dense_dyn = Tensor(shape=[None, None], dtype=mstype.int32)
|
|
alpha = Tensor(1, dtype=mstype.int32)
|
|
beta = Tensor(1, dtype=mstype.int32)
|
|
|
|
net.set_inputs(x1_indices_dyn, x1_values_dyn, x1_shape_dyn, x2_indices_dyn,
|
|
x2_values_dyn, x2_shape_dyn, x3_dense_dyn, alpha, beta)
|
|
|
|
x1_indices = Tensor(np.array([[0, 1], [0, 1]]), mstype.int64)
|
|
x1_values = Tensor(np.array([1, 2]), mstype.int32)
|
|
x1_shape = Tensor(np.array([3, 3]), mstype.int64)
|
|
x2_indices = Tensor(np.array([[0, 1], [2, 2]]), mstype.int64)
|
|
x2_values = Tensor(np.array([3, 4]), mstype.int32)
|
|
x2_shape = Tensor(np.array([3, 3]), mstype.int64)
|
|
x3_dense = Tensor(np.array([[1, 2, 3], [1, 3, 2], [3, 2, 1]]),
|
|
mstype.int32)
|
|
|
|
out = net(x1_indices, x1_values, x1_shape, x2_indices, x2_values, x2_shape,
|
|
x3_dense, alpha, beta)
|
|
expect_shapes = [(2, 8), (8,), (2,)]
|
|
for i in range(3):
|
|
assert out[i].asnumpy().shape == expect_shapes[i]
|
|
|
|
|
|
@pytest.mark.level0
|
|
@pytest.mark.platform_x86_gpu_training
|
|
@pytest.mark.env_onecard
|
|
def test_sspaddmm_input_int32():
|
|
"""
|
|
Feature: Sspaddmm gpu TEST.
|
|
Description: 2d int32 test case for Sspaddmm
|
|
Expectation: The value and shape of output are the expected values.
|
|
"""
|
|
context.set_context(mode=context.GRAPH_MODE, device_target="GPU")
|
|
|
|
x1_indices = Tensor(np.array([[0, 1], [0, 1]]), mstype.int32)
|
|
x1_values = Tensor(np.array([1, 2]), mstype.int32)
|
|
x1_shape = Tensor(np.array([3, 3]), mstype.int32)
|
|
x2_indices = Tensor(np.array([[0, 1], [2, 2]]), mstype.int32)
|
|
x2_values = Tensor(np.array([3, 4]), mstype.int32)
|
|
x2_shape = Tensor(np.array([3, 3]), mstype.int32)
|
|
x3_dense = Tensor(np.array([[1, 2, 3], [1, 3, 2], [3, 2, 1]]),
|
|
mstype.int32)
|
|
alpha = Tensor(np.array([1]), mstype.int32)
|
|
beta = Tensor(np.array([1]), mstype.int32)
|
|
net = SspaddmmNet()
|
|
y_indices, y_values, y_shape = net(x1_indices, x1_values, x1_shape,
|
|
x2_indices, x2_values, x2_shape,
|
|
x3_dense, alpha, beta)
|
|
y_indices_expect = np.array(
|
|
[[0, 1, 0, 0, 0, 1, 1, 1], [0, 1, 0, 1, 2, 0, 1, 2]], dtype=np.int64)
|
|
y_values_expect = np.array([1, 2, 9, 6, 3, 12, 8, 4], dtype=np.int32)
|
|
y_shape_expect = np.array([3, 3], dtype=np.int64)
|
|
|
|
assert np.allclose(y_indices.asnumpy(), y_indices_expect.astype(np.int64),
|
|
0.0001, 0.0001)
|
|
assert np.allclose(y_values.asnumpy(), y_values_expect.astype(np.int32),
|
|
0.0001, 0.0001)
|
|
assert np.allclose(y_shape.asnumpy(), y_shape_expect.astype(np.int64),
|
|
0.0001, 0.0001)
|
|
|
|
|
|
@pytest.mark.level0
|
|
@pytest.mark.platform_x86_gpu_training
|
|
@pytest.mark.env_onecard
|
|
def test_sspaddmm_input_int64():
|
|
"""
|
|
Feature: Sspaddmm gpu TEST.
|
|
Description: 2d int64 test case for Sspaddmm
|
|
Expectation: The value and shape of output are the expected values.
|
|
"""
|
|
context.set_context(mode=context.PYNATIVE_MODE, device_target="GPU")
|
|
|
|
x1_indices = Tensor(np.array([[0, 1], [0, 1]]), mstype.int32)
|
|
x1_values = Tensor(np.array([7, 6]), mstype.int32)
|
|
x1_shape = Tensor(np.array([3, 3]), mstype.int32)
|
|
x2_indices = Tensor(np.array([[0, 1], [2, 2]]), mstype.int32)
|
|
x2_values = Tensor(np.array([11, 23]), mstype.int32)
|
|
x2_shape = Tensor(np.array([3, 3]), mstype.int32)
|
|
x3_dense = Tensor(np.array([[1, 2, 3], [1, 3, 2], [3, 2, 1]]),
|
|
mstype.int32)
|
|
alpha = Tensor(np.array([2]), mstype.int32)
|
|
beta = Tensor(np.array([2]), mstype.int32)
|
|
net = SspaddmmNet()
|
|
y_indices, y_values, y_shape = net(x1_indices, x1_values, x1_shape,
|
|
x2_indices, x2_values, x2_shape,
|
|
x3_dense, alpha, beta)
|
|
y_indices_expect = np.array([[0, 1, 0, 0, 0, 1, 1, 1],
|
|
[0, 1, 0, 1, 2, 0, 1, 2]])
|
|
y_values_expect = np.array([14, 12, 66, 44, 22, 138, 92, 46])
|
|
y_shape_expect = np.array([3, 3])
|
|
|
|
assert np.allclose(y_indices.asnumpy(), y_indices_expect.astype(np.int64),
|
|
0.0001, 0.0001)
|
|
assert np.allclose(y_values.asnumpy(), y_values_expect.astype(np.int32),
|
|
0.0001, 0.0001)
|
|
assert np.allclose(y_shape.asnumpy(), y_shape_expect.astype(np.int64),
|
|
0.0001, 0.0001)
|