Files
2026-07-13 12:40:42 +08:00

638 lines
22 KiB
Python

# Copyright (c) 2022 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 numpy as np
import scipy.sparse as sp
from op_test import get_cuda_version, is_custom_device
import paddle
from paddle.base.framework import in_pir_mode
paddle.set_default_dtype('float64')
class TestMatmulSparseDense(unittest.TestCase):
# x: sparse, y: dense, out: dense
def check_result(self, x_shape, y_shape, format):
if len(x_shape) == 3:
mask = paddle.randint(0, 2, [x_shape[-2], x_shape[-1]])
else:
mask = paddle.randint(0, 2, x_shape)
origin_x = paddle.rand(x_shape) * mask.astype(
paddle.get_default_dtype()
)
origin_y = paddle.rand(y_shape)
dense_x = origin_x.detach()
dense_x.stop_gradient = False
dense_y = origin_y.detach()
dense_y.stop_gradient = False
dense_out = paddle.matmul(dense_x, dense_y)
if format == "coo":
sp_x = origin_x.detach().to_sparse_coo(len(x_shape))
else:
sp_x = origin_x.detach().to_sparse_csr()
sp_x.stop_gradient = False
sp_y = origin_y.detach()
sp_y.stop_gradient = False
sp_out = paddle.sparse.matmul(sp_x, sp_y)
np.testing.assert_allclose(
sp_out.numpy(), dense_out.numpy(), rtol=1e-05
)
if get_cuda_version() >= 11030:
dense_out.backward()
sp_out.backward()
np.testing.assert_allclose(
sp_x.grad.to_dense().numpy(),
(dense_x.grad * mask.astype(dense_x.dtype)).numpy(),
rtol=1e-05,
)
np.testing.assert_allclose(
sp_y.grad.numpy(), dense_y.grad.numpy(), rtol=1e-05
)
@unittest.skipIf(
not (paddle.is_compiled_with_cuda() or is_custom_device())
or paddle.is_compiled_with_rocm(),
"only support cuda",
)
def test_matmul_2d(self):
self.check_result([16, 12], [12, 10], 'coo')
self.check_result([16, 12], [12, 10], 'csr')
@unittest.skipIf(
not (paddle.is_compiled_with_cuda() or is_custom_device())
or get_cuda_version() < 11080,
"only support cuda>=11.8",
)
def test_matmul_3d(self):
self.check_result([8, 16, 12], [8, 12, 10], 'coo')
self.check_result([8, 16, 12], [8, 12, 10], 'csr')
class TestMatmulSparseSparseInt64Index(unittest.TestCase):
# x: sparse, y: sparse, out: sparse
def check_result(self, x_shape, y_shape, format):
origin_x = paddle.rand(x_shape)
origin_y = paddle.rand(y_shape)
dense_x = origin_x.detach()
dense_x.stop_gradient = False
dense_y = origin_y.detach()
dense_y.stop_gradient = False
dense_out = paddle.matmul(dense_x, dense_y)
if format == "coo":
sp_x = origin_x.detach().to_sparse_coo(len(x_shape))
sp_y = origin_y.detach().to_sparse_coo(len(y_shape))
else:
sp_x = origin_x.detach().to_sparse_csr()
sp_y = origin_y.detach().to_sparse_csr()
sp_x.stop_gradient = False
sp_y.stop_gradient = False
sp_out = paddle.sparse.matmul(sp_x, sp_y)
np.testing.assert_allclose(
sp_out.to_dense().numpy(), dense_out.numpy(), rtol=1e-05
)
dense_out.backward()
sp_out.backward()
np.testing.assert_allclose(
sp_x.grad.to_dense().numpy(),
dense_x.grad.numpy(),
rtol=1e-05,
)
np.testing.assert_allclose(
sp_y.grad.to_dense().numpy(), dense_y.grad.numpy(), rtol=1e-05
)
@unittest.skipIf(
not (paddle.is_compiled_with_cuda() or is_custom_device())
or paddle.is_compiled_with_rocm(),
"only support cuda",
)
def test_matmul_2d(self):
self.check_result([16, 12], [12, 10], 'coo')
self.check_result([16, 12], [12, 10], 'csr')
@unittest.skipIf(
not (paddle.is_compiled_with_cuda() or is_custom_device())
or paddle.is_compiled_with_rocm(),
"only support cuda",
)
def test_matmul_3d(self):
self.check_result([8, 16, 12], [8, 12, 10], 'coo')
self.check_result([8, 16, 12], [8, 12, 10], 'csr')
class TestMatmulSparseSparseInt32Index(unittest.TestCase):
# x: sparse, y: sparse, out: sparse
def check_result(self, x_shape, y_shape, format):
origin_x = paddle.rand(x_shape)
origin_y = paddle.rand(y_shape)
dense_x = origin_x.detach()
dense_x.stop_gradient = False
dense_y = origin_y.detach()
dense_y.stop_gradient = False
dense_out = paddle.matmul(dense_x, dense_y)
if format == "coo":
sp_x = origin_x.detach().to_sparse_coo(len(x_shape))
# cast to 32-bit index.
sp_x_indices = paddle.cast(sp_x.indices(), "int32")
sp_x = paddle.sparse.sparse_coo_tensor(
sp_x_indices, sp_x.values(), sp_x.shape
)
sp_y = origin_y.detach().to_sparse_coo(len(y_shape))
# cast to 32-bit index.
sp_y_indices = paddle.cast(sp_y.indices(), "int32")
sp_y = paddle.sparse.sparse_coo_tensor(
sp_y_indices, sp_y.values(), sp_y.shape
)
else:
sp_x = origin_x.detach().to_sparse_csr()
# cast to 32-bit index.
sp_x_crows = paddle.cast(sp_x.crows(), "int32")
sp_x_cols = paddle.cast(sp_x.cols(), "int32")
sp_x = paddle.sparse.sparse_csr_tensor(
sp_x_crows, sp_x_cols, sp_x.values(), sp_x.shape
)
sp_y = origin_y.detach().to_sparse_csr()
# cast to 32-bit index.
sp_y_crows = paddle.cast(sp_y.crows(), "int32")
sp_y_cols = paddle.cast(sp_y.cols(), "int32")
sp_y = paddle.sparse.sparse_csr_tensor(
sp_y_crows, sp_y_cols, sp_y.values(), sp_y.shape
)
sp_x.stop_gradient = False
sp_y.stop_gradient = False
sp_out = paddle.sparse.matmul(sp_x, sp_y)
np.testing.assert_allclose(
sp_out.to_dense().numpy(), dense_out.numpy(), rtol=1e-05
)
dense_out.backward()
sp_out.backward()
np.testing.assert_allclose(
sp_x.grad.to_dense().numpy(),
dense_x.grad.numpy(),
rtol=1e-05,
)
np.testing.assert_allclose(
sp_y.grad.to_dense().numpy(), dense_y.grad.numpy(), rtol=1e-05
)
@unittest.skipIf(
not (paddle.is_compiled_with_cuda() or is_custom_device())
or paddle.is_compiled_with_rocm(),
"only support cuda",
)
def test_matmul_2d(self):
self.check_result([16, 12], [12, 10], 'coo')
self.check_result([16, 12], [12, 10], 'csr')
@unittest.skipIf(
not (paddle.is_compiled_with_cuda() or is_custom_device())
or paddle.is_compiled_with_rocm(),
"only support cuda",
)
def test_matmul_3d(self):
self.check_result([8, 16, 12], [8, 12, 10], 'coo')
self.check_result([8, 16, 12], [8, 12, 10], 'csr')
class TestMaskedMatmul(unittest.TestCase):
# x: dense, y: dense, out: sparse_`csr
@unittest.skipIf(
not (paddle.is_compiled_with_cuda() or is_custom_device())
or get_cuda_version() < 11030,
"only support on cuda>=11.3",
)
def test_masked_matmul_2d(self):
np_mask = np.random.rand(10, 6) < 0.2
np_x = np.random.rand(10, 12)
np_y = np.random.rand(12, 6)
np_out = sp.csr_matrix(np.matmul(np_x, np_y) * np_mask)
np_out_grad = sp.csr_matrix(np.ones([10, 6]) * np_mask)
# dx(dense) = dout(csr) * y'(dense)
np_x_grad = np_out_grad @ np_y.transpose(1, 0)
# dy(dense) = x'(dense) * dout(csr) -> dy'(dense) = dout'(csr) * x(dense)
np_y_grad = (np_out_grad.transpose() @ np_x).transpose(1, 0)
x = paddle.to_tensor(np_x, stop_gradient=False)
y = paddle.to_tensor(np_y, stop_gradient=False)
mask = paddle.to_tensor(np.ones([10, 6]) * np_mask).to_sparse_csr()
out = paddle.sparse.masked_matmul(x, y, mask)
np.testing.assert_allclose(
np_out.indptr, out.crows().numpy(), rtol=1e-05
)
np.testing.assert_allclose(
np_out.indices, out.cols().numpy(), rtol=1e-05
)
np.testing.assert_allclose(
np_out.data, out.values().numpy(), rtol=1e-05
)
out.backward()
np.testing.assert_allclose(out.is_sparse_csr(), True, rtol=1e-05)
np.testing.assert_allclose(np_x_grad, x.grad.numpy(), rtol=1e-05)
np.testing.assert_allclose(np_y_grad, y.grad.numpy(), rtol=1e-05)
@unittest.skipIf(
not (paddle.is_compiled_with_cuda() or is_custom_device())
or get_cuda_version() < 11080,
"only support on cuda>=11.8",
)
def test_masked_matmul_3d(self):
paddle.set_default_dtype('float32')
origin_x = paddle.rand([16, 16, 12])
mask = paddle.randint(0, 2, [16, 12])
origin_x = origin_x * mask.astype('float32')
origin_y = paddle.rand([16, 12, 10])
dense_x = origin_x.detach()
dense_x.stop_gradient = False
dense_y = origin_y.detach()
dense_y.stop_gradient = False
dense_out = paddle.matmul(dense_x, dense_y)
dense_out.backward()
sp_x = origin_x.detach().to_sparse_csr()
sp_x.stop_gradient = False
sp_y = origin_y.detach()
sp_y.stop_gradient = False
sp_out = paddle.sparse.matmul(sp_x, sp_y)
sp_out.backward()
np.testing.assert_allclose(
sp_out.numpy(), dense_out.numpy(), rtol=1e-05
)
np.testing.assert_allclose(
sp_x.grad.to_dense().numpy(),
(dense_x.grad * mask.astype('float32')).numpy(),
rtol=1e-05,
)
np.testing.assert_allclose(
sp_y.grad.numpy(), dense_y.grad.numpy(), rtol=1e-05
)
class TestMatmulSparseDenseStatic(unittest.TestCase):
# x: sparse, y: dense, out: dense
def check_result(self, x_shape, y_shape):
# only support sparse_coo_tensor in static graph
if len(x_shape) == 3:
mask = paddle.randint(0, 2, [x_shape[-2], x_shape[-1]])
else:
mask = paddle.randint(0, 2, x_shape)
origin_x = paddle.rand(x_shape) * mask.astype(
paddle.get_default_dtype()
)
origin_y = paddle.rand(y_shape)
dense_x = origin_x.detach()
dense_y = origin_y.detach()
dense_out = paddle.matmul(dense_x, dense_y)
indices_data, values_data = (
origin_x.detach().to_sparse_coo(len(x_shape)).indices(),
origin_x.detach().to_sparse_coo(len(x_shape)).values(),
)
paddle.enable_static()
with paddle.static.program_guard(
paddle.static.Program(), paddle.static.Program()
):
indices = paddle.static.data(
name='indices',
shape=indices_data.shape,
dtype=indices_data.dtype,
)
values = paddle.static.data(
name='values',
shape=values_data.shape,
dtype=values_data.dtype,
)
sp_x = paddle.sparse.sparse_coo_tensor(
indices,
values,
shape=origin_x.shape,
dtype=origin_x.dtype,
)
sp_y = paddle.static.data(
name='sp_y',
shape=origin_y.shape,
dtype=origin_y.dtype,
)
sp_out = paddle.sparse.matmul(sp_x, sp_y)
exe = paddle.static.Executor()
fetch = exe.run(
feed={
'indices': indices_data.numpy(),
'values': values_data.numpy(),
'sp_y': origin_y.detach().numpy(),
},
fetch_list=[sp_out],
return_numpy=False,
)
sp_out = fetch[0]
np.testing.assert_allclose(
sp_out.numpy(), dense_out.numpy(), rtol=1e-05
)
paddle.disable_static()
@unittest.skipIf(
not (paddle.is_compiled_with_cuda() or is_custom_device())
or paddle.is_compiled_with_rocm(),
"only support cuda",
)
def test_matmul_2d(self):
if in_pir_mode():
self.check_result([16, 12], [12, 10])
@unittest.skipIf(
not (paddle.is_compiled_with_cuda() or is_custom_device())
or get_cuda_version() < 11080,
"only support cuda>=11.8",
)
def test_matmul_3d(self):
if in_pir_mode():
self.check_result([8, 16, 12], [8, 12, 10])
class TestMatmulSparseSparseStatic(unittest.TestCase):
'''
only support sparse_coo_tensor in static graph
'''
# x: sparse, y: sparse, out: sparse
def check_result(self, x_shape, y_shape):
origin_x = paddle.rand(x_shape)
origin_y = paddle.rand(y_shape)
dense_x = origin_x.detach()
dense_y = origin_y.detach()
dense_out = paddle.matmul(dense_x, dense_y)
x_indices_data, x_values_data = (
origin_x.detach().to_sparse_coo(len(x_shape)).indices(),
origin_x.detach().to_sparse_coo(len(x_shape)).values(),
)
y_indices_data, y_values_data = (
origin_y.detach().to_sparse_coo(len(y_shape)).indices(),
origin_y.detach().to_sparse_coo(len(y_shape)).values(),
)
paddle.enable_static()
with paddle.static.program_guard(
paddle.static.Program(), paddle.static.Program()
):
x_indices = paddle.static.data(
name='x_indices',
shape=x_indices_data.shape,
dtype=x_indices_data.dtype,
)
x_values = paddle.static.data(
name='x_values',
shape=x_values_data.shape,
dtype=x_values_data.dtype,
)
sp_x = paddle.sparse.sparse_coo_tensor(
x_indices,
x_values,
shape=origin_x.shape,
dtype=origin_x.dtype,
)
y_indices = paddle.static.data(
name='y_indices',
shape=y_indices_data.shape,
dtype=y_indices_data.dtype,
)
y_values = paddle.static.data(
name='y_values',
shape=y_values_data.shape,
dtype=y_values_data.dtype,
)
sp_y = paddle.sparse.sparse_coo_tensor(
y_indices,
y_values,
shape=origin_y.shape,
dtype=origin_y.dtype,
)
sp_out = paddle.sparse.matmul(sp_x, sp_y)
exe = paddle.static.Executor()
fetch = exe.run(
feed={
'x_indices': x_indices_data.numpy(),
'x_values': x_values_data.numpy(),
'y_indices': y_indices_data.numpy(),
'y_values': y_values_data.numpy(),
},
fetch_list=[sp_out],
return_numpy=False,
)
sp_out = fetch[0]
np.testing.assert_allclose(
sp_out.to_dense().numpy(), dense_out.numpy(), rtol=1e-05
)
paddle.disable_static()
@unittest.skipIf(
not (paddle.is_compiled_with_cuda() or is_custom_device())
or paddle.is_compiled_with_rocm(),
"only support cuda",
)
def test_matmul_2d(self):
if in_pir_mode():
self.check_result([16, 12], [12, 10])
@unittest.skipIf(
not (paddle.is_compiled_with_cuda() or is_custom_device())
or paddle.is_compiled_with_rocm(),
"only support cuda",
)
def test_matmul_3d(self):
if in_pir_mode():
self.check_result([8, 16, 12], [8, 12, 10])
class TestMaskedMatmulStatic(unittest.TestCase):
'''
only support sparse_csr_tensor in static graph
'''
# x: dense, y: dense, out: sparse_csr
@unittest.skipIf(
not (paddle.is_compiled_with_cuda() or is_custom_device())
or get_cuda_version() < 11030,
"only support on cuda>=11.3",
)
def test_masked_matmul_2d(self):
if in_pir_mode():
np_mask = np.random.rand(10, 6) < 0.2
np_x = np.random.rand(10, 12)
np_y = np.random.rand(12, 6)
x = paddle.to_tensor(np_x)
y = paddle.to_tensor(np_y)
mask = paddle.to_tensor(np.ones([10, 6]) * np_mask).to_sparse_coo(
len(np_mask.shape)
)
out = paddle.sparse.masked_matmul(x, y, mask)
indices_data, values_data = (
mask.indices(),
mask.values(),
)
paddle.enable_static()
with paddle.static.program_guard(
paddle.static.Program(), paddle.static.Program()
):
indices = paddle.static.data(
name='indices',
shape=indices_data.shape,
dtype=indices_data.dtype,
)
values = paddle.static.data(
name='values',
shape=values_data.shape,
dtype=values_data.dtype,
)
sp_mask = paddle.sparse.sparse_coo_tensor(
indices,
values,
shape=mask.shape,
dtype=mask.dtype,
)
sp_x = paddle.static.data(
name='x',
shape=x.shape,
dtype=x.dtype,
)
sp_y = paddle.static.data(
name='y',
shape=y.shape,
dtype=y.dtype,
)
out = paddle.sparse.masked_matmul(sp_x, sp_y, sp_mask)
exe = paddle.static.Executor()
fetch = exe.run(
feed={
'indices': indices_data.numpy(),
'values': values_data.numpy(),
'x': x.numpy(),
'y': y.numpy(),
},
fetch_list=[out],
return_numpy=False,
)
sp_out = fetch[0]
np.testing.assert_allclose(
sp_out.to_dense().numpy(),
out.to_dense().numpy(),
rtol=1e-05,
)
paddle.disable_static()
@unittest.skipIf(
not (paddle.is_compiled_with_cuda() or is_custom_device())
or get_cuda_version() < 11080,
"only support on cuda>=11.8",
)
def test_masked_matmul_3d(self):
if in_pir_mode():
paddle.set_default_dtype('float32')
origin_x = paddle.rand([16, 16, 12])
mask = paddle.randint(0, 2, [16, 12])
origin_x = origin_x * mask.astype('float32')
origin_y = paddle.rand([16, 12, 10])
x = origin_x.detach()
y = origin_y.detach()
mask = paddle.to_tensor(np.ones([16, 12]) * mask).to_sparse_coo(
len(mask.shape)
)
out = paddle.sparse.masked_matmul(x, y, mask)
indices_data, values_data = (
mask.indices(),
mask.values(),
)
paddle.enable_static()
with paddle.static.program_guard(
paddle.static.Program(), paddle.static.Program()
):
indices = paddle.static.data(
name='indices',
shape=indices_data.shape,
dtype=indices_data.dtype,
)
values = paddle.static.data(
name='values',
shape=values_data.shape,
dtype=values_data.dtype,
)
sp_mask = paddle.sparse.sparse_coo_tensor(
indices,
values,
shape=mask.shape,
dtype=mask.dtype,
)
sp_x = paddle.static.data(
name='x',
shape=origin_x.shape,
dtype=origin_x.dtype,
)
sp_y = paddle.static.data(
name='y',
shape=origin_y.shape,
dtype=origin_y.dtype,
)
out = paddle.sparse.masked_matmul(sp_x, sp_y, sp_mask)
exe = paddle.static.Executor()
fetch = exe.run(
feed={
'indices': indices_data.numpy(),
'values': values_data.numpy(),
'x': origin_x.numpy(),
'y': origin_y.numpy(),
},
fetch_list=[out],
return_numpy=False,
)
sp_out = fetch[0]
np.testing.assert_allclose(
sp_out.to_dense().numpy(),
out.to_dense().numpy(),
rtol=1e-05,
)
paddle.disable_static()
if __name__ == "__main__":
unittest.main()