489 lines
18 KiB
Python
489 lines
18 KiB
Python
# Copyright (c) 2020 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 sys
|
|
import unittest
|
|
|
|
import numpy as np
|
|
from op_test import get_device_place, is_custom_device
|
|
|
|
sys.path.append("../../legacy_test")
|
|
from test_sum_op import TestReduceOPTensorAxisBase
|
|
from utils import dygraph_guard, static_guard
|
|
|
|
import paddle
|
|
from paddle.framework import core
|
|
|
|
|
|
class TestProdOp(unittest.TestCase):
|
|
def setUp(self):
|
|
self.input = np.random.random(size=(10, 10, 5)).astype(np.float32)
|
|
|
|
def run_imperative(self, place):
|
|
input = paddle.to_tensor(self.input, place=place)
|
|
out = paddle.prod(input)
|
|
expected_result = np.prod(self.input)
|
|
np.testing.assert_allclose(out.numpy(), expected_result, rtol=1e-05)
|
|
|
|
out = paddle.prod(input, axis=1)
|
|
expected_result = np.prod(self.input, axis=1)
|
|
np.testing.assert_allclose(out.numpy(), expected_result, rtol=1e-05)
|
|
|
|
out = paddle.prod(input, axis=-1)
|
|
expected_result = np.prod(self.input, axis=-1)
|
|
np.testing.assert_allclose(out.numpy(), expected_result, rtol=1e-05)
|
|
|
|
out = paddle.prod(input, axis=[0, 1])
|
|
expected_result = np.prod(self.input, axis=(0, 1))
|
|
np.testing.assert_allclose(
|
|
out.numpy(), expected_result, rtol=1e-05, atol=1e-8
|
|
)
|
|
|
|
out = paddle.prod(input, axis=1, keepdim=True)
|
|
expected_result = np.prod(self.input, axis=1, keepdims=True)
|
|
np.testing.assert_allclose(out.numpy(), expected_result, rtol=1e-05)
|
|
|
|
out = paddle.prod(input, axis=1, dtype='int64')
|
|
expected_result = np.prod(self.input, axis=1, dtype=np.int64)
|
|
np.testing.assert_allclose(out.numpy(), expected_result, rtol=1e-05)
|
|
|
|
out = paddle.prod(input, axis=1, keepdim=True, dtype='int64')
|
|
expected_result = np.prod(
|
|
self.input, axis=1, keepdims=True, dtype=np.int64
|
|
)
|
|
np.testing.assert_allclose(out.numpy(), expected_result, rtol=1e-05)
|
|
|
|
def run_static(self, use_gpu=False):
|
|
with paddle.static.program_guard(paddle.static.Program()):
|
|
input = paddle.static.data(
|
|
name='input', shape=[10, 10, 5], dtype='float32'
|
|
)
|
|
result0 = paddle.prod(input)
|
|
result1 = paddle.prod(input, axis=1)
|
|
result2 = paddle.prod(input, axis=-1)
|
|
result3 = paddle.prod(input, axis=[0, 1])
|
|
result4 = paddle.prod(input, axis=1, keepdim=True)
|
|
result5 = paddle.prod(input, axis=1, dtype='int64')
|
|
result6 = paddle.prod(input, axis=1, keepdim=True, dtype='int64')
|
|
|
|
place = get_device_place() if use_gpu else paddle.CPUPlace()
|
|
exe = paddle.static.Executor(place)
|
|
exe.run(paddle.static.default_startup_program())
|
|
static_result = exe.run(
|
|
feed={"input": self.input},
|
|
fetch_list=[
|
|
result0,
|
|
result1,
|
|
result2,
|
|
result3,
|
|
result4,
|
|
result5,
|
|
result6,
|
|
],
|
|
)
|
|
|
|
expected_result = np.prod(self.input)
|
|
np.testing.assert_allclose(
|
|
static_result[0], expected_result, rtol=1e-05
|
|
)
|
|
expected_result = np.prod(self.input, axis=1)
|
|
np.testing.assert_allclose(
|
|
static_result[1], expected_result, rtol=1e-05
|
|
)
|
|
expected_result = np.prod(self.input, axis=-1)
|
|
np.testing.assert_allclose(
|
|
static_result[2], expected_result, rtol=1e-05
|
|
)
|
|
expected_result = np.prod(self.input, axis=(0, 1))
|
|
np.testing.assert_allclose(
|
|
static_result[3], expected_result, rtol=1e-05, atol=1e-8
|
|
)
|
|
expected_result = np.prod(self.input, axis=1, keepdims=True)
|
|
np.testing.assert_allclose(
|
|
static_result[4], expected_result, rtol=1e-05
|
|
)
|
|
expected_result = np.prod(self.input, axis=1, dtype=np.int64)
|
|
np.testing.assert_allclose(
|
|
static_result[5], expected_result, rtol=1e-05
|
|
)
|
|
expected_result = np.prod(
|
|
self.input, axis=1, keepdims=True, dtype=np.int64
|
|
)
|
|
np.testing.assert_allclose(
|
|
static_result[6], expected_result, rtol=1e-05
|
|
)
|
|
|
|
def test_cpu(self):
|
|
with dygraph_guard():
|
|
self.run_imperative(place=paddle.CPUPlace())
|
|
with static_guard():
|
|
self.run_static()
|
|
|
|
def test_gpu(self):
|
|
if not (paddle.base.core.is_compiled_with_cuda() or is_custom_device()):
|
|
return
|
|
with dygraph_guard():
|
|
self.run_imperative(place=get_device_place())
|
|
with static_guard():
|
|
self.run_static()
|
|
|
|
|
|
@unittest.skipIf(
|
|
core.is_compiled_with_xpu(),
|
|
"Skip XPU for complex dtype is not fully supported",
|
|
)
|
|
class TestProdComplexOp(TestProdOp):
|
|
def setUp(self):
|
|
real = np.random.random(size=(10, 10, 5)).astype(np.float32)
|
|
imag = np.random.random(size=(10, 10, 5)).astype(np.float32)
|
|
self.input = real + 1j * imag
|
|
|
|
def run_imperative(self, place):
|
|
input = paddle.to_tensor(self.input, place=place)
|
|
out = paddle.prod(input)
|
|
expected_result = np.prod(self.input)
|
|
np.testing.assert_allclose(out.numpy(), expected_result, rtol=1e-05)
|
|
|
|
out = paddle.prod(input, axis=1)
|
|
expected_result = np.prod(self.input, axis=1)
|
|
np.testing.assert_allclose(out.numpy(), expected_result, rtol=1e-05)
|
|
|
|
out = paddle.prod(input, axis=[0, 1])
|
|
expected_result = np.prod(self.input, axis=(0, 1))
|
|
np.testing.assert_allclose(
|
|
out.numpy(), expected_result, rtol=1e-05, atol=1e-8
|
|
)
|
|
|
|
out = paddle.prod(input, axis=1, keepdim=True)
|
|
expected_result = np.prod(self.input, axis=1, keepdims=True)
|
|
np.testing.assert_allclose(out.numpy(), expected_result, rtol=1e-05)
|
|
|
|
def run_static(self, use_gpu=False):
|
|
with paddle.static.program_guard(paddle.static.Program()):
|
|
input = paddle.static.data(
|
|
name='input', shape=[10, 10, 5], dtype='complex64'
|
|
)
|
|
result0 = paddle.prod(input)
|
|
result1 = paddle.prod(input, axis=1)
|
|
result2 = paddle.prod(input, axis=-1)
|
|
result3 = paddle.prod(input, axis=[0, 1])
|
|
result4 = paddle.prod(input, axis=1, keepdim=True)
|
|
|
|
place = get_device_place() if use_gpu else paddle.CPUPlace()
|
|
exe = paddle.static.Executor(place)
|
|
exe.run(paddle.static.default_startup_program())
|
|
static_complex_result = exe.run(
|
|
feed={"input": self.input},
|
|
fetch_list=[
|
|
result0,
|
|
result1,
|
|
result2,
|
|
result3,
|
|
result4,
|
|
],
|
|
)
|
|
|
|
expected_result = np.prod(self.input)
|
|
np.testing.assert_allclose(
|
|
static_complex_result[0], expected_result, rtol=1e-05
|
|
)
|
|
expected_result = np.prod(self.input, axis=1)
|
|
np.testing.assert_allclose(
|
|
static_complex_result[1], expected_result, rtol=1e-05
|
|
)
|
|
expected_result = np.prod(self.input, axis=-1)
|
|
np.testing.assert_allclose(
|
|
static_complex_result[2], expected_result, rtol=1e-05
|
|
)
|
|
expected_result = np.prod(self.input, axis=(0, 1))
|
|
np.testing.assert_allclose(
|
|
static_complex_result[3], expected_result, rtol=1e-05, atol=1e-8
|
|
)
|
|
expected_result = np.prod(self.input, axis=1, keepdims=True)
|
|
np.testing.assert_allclose(
|
|
static_complex_result[4], expected_result, rtol=1e-05
|
|
)
|
|
|
|
def test_cpu(self):
|
|
with dygraph_guard():
|
|
self.run_imperative(place=paddle.CPUPlace())
|
|
with static_guard():
|
|
self.run_static()
|
|
|
|
def test_gpu(self):
|
|
if not (paddle.base.core.is_compiled_with_cuda() or is_custom_device()):
|
|
return
|
|
with dygraph_guard():
|
|
self.run_imperative(place=get_device_place())
|
|
with static_guard():
|
|
self.run_static()
|
|
|
|
|
|
class TestProdOpError(unittest.TestCase):
|
|
def test_error(self):
|
|
with (
|
|
static_guard(),
|
|
paddle.static.program_guard(
|
|
paddle.static.Program(), paddle.static.Program()
|
|
),
|
|
):
|
|
x = paddle.static.data(name='x', shape=[2, 2, 4], dtype='float32')
|
|
bool_x = paddle.static.data(
|
|
name='bool_x', shape=[2, 2, 4], dtype='bool'
|
|
)
|
|
# The argument x should be a Tensor
|
|
self.assertRaises(TypeError, paddle.prod, [1])
|
|
|
|
# The data type of x should be float32, float64, int32, int64
|
|
self.assertRaises(TypeError, paddle.prod, bool_x)
|
|
|
|
# The argument axis's type should be int ,list or tuple
|
|
self.assertRaises(TypeError, paddle.prod, x, 1.5)
|
|
|
|
# The argument dtype of prod_op should be float32, float64, int32 or int64.
|
|
self.assertRaises(TypeError, paddle.prod, x, 'bool')
|
|
|
|
|
|
class TestProdWithTensorAxis1(TestReduceOPTensorAxisBase):
|
|
def init_data(self):
|
|
self.pd_api = paddle.prod
|
|
self.np_api = np.prod
|
|
self.x = paddle.randn([10, 5, 9, 9], dtype='float64')
|
|
self.np_axis = np.array([1, 2], dtype='int64')
|
|
self.tensor_axis = paddle.to_tensor([1, 2], dtype='int64')
|
|
|
|
|
|
class TestProdWithTensorAxis2(TestReduceOPTensorAxisBase):
|
|
def init_data(self):
|
|
self.pd_api = paddle.prod
|
|
self.np_api = np.prod
|
|
self.x = paddle.randn([10, 10, 9, 9], dtype='float64')
|
|
self.np_axis = np.array([0, 1, 2], dtype='int64')
|
|
self.tensor_axis = [
|
|
0,
|
|
paddle.to_tensor([1], 'int64'),
|
|
paddle.to_tensor([2], 'int64'),
|
|
]
|
|
|
|
|
|
class TestProdOp_ZeroSize(unittest.TestCase):
|
|
def setUp(self):
|
|
self.input = np.random.random(size=(10, 0, 5)).astype(np.float32)
|
|
|
|
def run_imperative(self, place):
|
|
input = paddle.to_tensor(self.input, place=place)
|
|
input.stop_gradient = False
|
|
out = paddle.prod(input)
|
|
expected_result = np.prod(self.input)
|
|
np.testing.assert_allclose(out.numpy(), expected_result, rtol=1e-05)
|
|
out.sum().backward()
|
|
np.testing.assert_allclose(input.grad.shape, input.shape)
|
|
|
|
def test_cpu(self):
|
|
with dygraph_guard():
|
|
self.run_imperative(place=paddle.CPUPlace())
|
|
|
|
def test_gpu(self):
|
|
if not (paddle.base.core.is_compiled_with_cuda() or is_custom_device()):
|
|
return
|
|
with dygraph_guard():
|
|
self.run_imperative(place=get_device_place())
|
|
|
|
|
|
class TestProdOp_ZeroSize2(TestProdOp_ZeroSize):
|
|
def setUp(self):
|
|
self.input = np.random.random(size=(10, 1, 5)).astype(np.float32)
|
|
|
|
def run_imperative(self, place):
|
|
input = paddle.to_tensor(self.input, place=place)
|
|
out = paddle.prod(input, paddle.randn([0]).astype(paddle.int32))
|
|
np.testing.assert_allclose(out.numpy(), input.numpy())
|
|
|
|
|
|
class TestProdAliasOp(unittest.TestCase):
|
|
def setUp(self):
|
|
self.input = np.random.random(size=(10, 10, 5)).astype(np.float32)
|
|
|
|
def run_imperative(self, place):
|
|
input = paddle.to_tensor(self.input, place=place)
|
|
out = paddle.prod(input=input)
|
|
expected_result = np.prod(self.input)
|
|
np.testing.assert_allclose(out.numpy(), expected_result, rtol=1e-05)
|
|
|
|
out = paddle.prod(input, dim=1)
|
|
expected_result = np.prod(self.input, axis=1)
|
|
np.testing.assert_allclose(out.numpy(), expected_result, rtol=1e-05)
|
|
|
|
out = paddle.prod(input=input, dim=-1)
|
|
expected_result = np.prod(self.input, axis=-1)
|
|
np.testing.assert_allclose(out.numpy(), expected_result, rtol=1e-05)
|
|
|
|
out = paddle.prod(input, dim=[0, 1])
|
|
expected_result = np.prod(self.input, axis=(0, 1))
|
|
np.testing.assert_allclose(
|
|
out.numpy(), expected_result, rtol=1e-05, atol=1e-8
|
|
)
|
|
|
|
out = paddle.prod(input, dim=1, keepdim=True)
|
|
expected_result = np.prod(self.input, axis=1, keepdims=True)
|
|
np.testing.assert_allclose(out.numpy(), expected_result, rtol=1e-05)
|
|
|
|
out = paddle.prod(input=input, dim=1, dtype='int64')
|
|
expected_result = np.prod(self.input, axis=1, dtype=np.int64)
|
|
np.testing.assert_allclose(out.numpy(), expected_result, rtol=1e-05)
|
|
|
|
out = paddle.prod(input=input, dim=1, keepdim=True, dtype='int64')
|
|
expected_result = np.prod(
|
|
self.input, axis=1, keepdims=True, dtype=np.int64
|
|
)
|
|
np.testing.assert_allclose(out.numpy(), expected_result, rtol=1e-05)
|
|
|
|
paddle_out2 = paddle.empty(expected_result.shape, dtype='int64')
|
|
paddle_out1 = paddle.prod(
|
|
input=input, dim=1, keepdim=True, dtype='int64', out=paddle_out2
|
|
)
|
|
np.testing.assert_allclose(
|
|
paddle_out1.numpy(), expected_result, rtol=1e-05
|
|
)
|
|
np.testing.assert_allclose(
|
|
paddle_out2.numpy(), expected_result, rtol=1e-05
|
|
)
|
|
|
|
def run_static(self, use_gpu=False):
|
|
with paddle.static.program_guard(paddle.static.Program()):
|
|
input = paddle.static.data(
|
|
name='input', shape=[10, 10, 5], dtype='float32'
|
|
)
|
|
expected_result = np.prod(self.input)
|
|
|
|
result0 = paddle.prod(input=input)
|
|
result1 = paddle.prod(input, dim=1)
|
|
result2 = paddle.prod(input=input, dim=-1)
|
|
result3 = paddle.prod(input, dim=[0, 1])
|
|
result4 = paddle.prod(input, dim=1, keepdim=True)
|
|
result5 = paddle.prod(input=input, dim=1, dtype='int64')
|
|
result6 = paddle.prod(input, dim=1, keepdim=True, dtype='int64')
|
|
|
|
result7 = paddle.zeros(shape=expected_result.shape, dtype="int64")
|
|
paddle.prod(input, dim=1, keepdim=True, dtype='int64', out=result7)
|
|
|
|
result8 = paddle.zeros(shape=expected_result.shape, dtype="int64")
|
|
result9 = paddle.prod(
|
|
input, dim=1, keepdim=True, dtype='int64', out=result8
|
|
)
|
|
|
|
place = get_device_place() if use_gpu else paddle.CPUPlace()
|
|
exe = paddle.static.Executor(place)
|
|
exe.run(paddle.static.default_startup_program())
|
|
static_result = exe.run(
|
|
feed={"input": self.input},
|
|
fetch_list=[
|
|
result0,
|
|
result1,
|
|
result2,
|
|
result3,
|
|
result4,
|
|
result5,
|
|
result6,
|
|
result7,
|
|
result8,
|
|
result9,
|
|
],
|
|
)
|
|
|
|
np.testing.assert_allclose(
|
|
static_result[0], expected_result, rtol=1e-05
|
|
)
|
|
expected_result = np.prod(self.input, axis=1)
|
|
np.testing.assert_allclose(
|
|
static_result[1], expected_result, rtol=1e-05
|
|
)
|
|
expected_result = np.prod(self.input, axis=-1)
|
|
np.testing.assert_allclose(
|
|
static_result[2], expected_result, rtol=1e-05
|
|
)
|
|
expected_result = np.prod(self.input, axis=(0, 1))
|
|
np.testing.assert_allclose(
|
|
static_result[3], expected_result, rtol=1e-05, atol=1e-8
|
|
)
|
|
expected_result = np.prod(self.input, axis=1, keepdims=True)
|
|
np.testing.assert_allclose(
|
|
static_result[4], expected_result, rtol=1e-05
|
|
)
|
|
expected_result = np.prod(self.input, axis=1, dtype=np.int64)
|
|
np.testing.assert_allclose(
|
|
static_result[5], expected_result, rtol=1e-05
|
|
)
|
|
expected_result = np.prod(
|
|
self.input, axis=1, keepdims=True, dtype=np.int64
|
|
)
|
|
np.testing.assert_allclose(
|
|
static_result[6], expected_result, rtol=1e-05
|
|
)
|
|
np.testing.assert_allclose(
|
|
static_result[7], expected_result, rtol=1e-05
|
|
)
|
|
np.testing.assert_allclose(
|
|
static_result[8], expected_result, rtol=1e-05
|
|
)
|
|
np.testing.assert_allclose(
|
|
static_result[9], expected_result, rtol=1e-05
|
|
)
|
|
|
|
def test_cpu(self):
|
|
with dygraph_guard():
|
|
self.run_imperative(place=paddle.CPUPlace())
|
|
with static_guard():
|
|
self.run_static()
|
|
|
|
def test_gpu(self):
|
|
if not (paddle.base.core.is_compiled_with_cuda() or is_custom_device()):
|
|
return
|
|
with dygraph_guard():
|
|
self.run_imperative(place=get_device_place())
|
|
with static_guard():
|
|
self.run_static()
|
|
|
|
def test_tensor_prod(self):
|
|
"""x.prod(axis=1) is equivalent to x.prod(dim=1)"""
|
|
axis_cases = [0, 1, -1]
|
|
|
|
def run_test_cases(place):
|
|
"""Helper function to run test cases on specified device."""
|
|
for param_alias in ["axis", "dim"]:
|
|
for axis in axis_cases:
|
|
input_tensor = paddle.to_tensor(self.input, place=place)
|
|
kwargs = {param_alias: axis}
|
|
|
|
result = input_tensor.prod(**kwargs)
|
|
expected = np.prod(self.input, axis=axis)
|
|
np.testing.assert_allclose(
|
|
(
|
|
result.numpy()
|
|
if place.is_cpu_place()
|
|
else result.cpu().numpy()
|
|
),
|
|
expected,
|
|
rtol=1e-05,
|
|
)
|
|
|
|
with dygraph_guard():
|
|
run_test_cases(paddle.CPUPlace())
|
|
|
|
if paddle.base.core.is_compiled_with_cuda() or is_custom_device():
|
|
run_test_cases(get_device_place())
|
|
|
|
|
|
if __name__ == "__main__":
|
|
unittest.main()
|