You can not select more than 25 topics
Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
957 lines
34 KiB
957 lines
34 KiB
# Copyright 2019 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 mindspore as ms
|
|
import mindspore.nn as nn
|
|
from mindspore import Tensor
|
|
from mindspore import context
|
|
from mindspore.common.api import _executor
|
|
from mindspore.ops import composite as C
|
|
from mindspore.ops import operations as P
|
|
from tests.ut.python.ops.test_math_ops import VirtualLoss
|
|
|
|
|
|
grad_all = C.GradOperation(get_all=True)
|
|
|
|
|
|
class NetWithLoss(nn.Cell):
|
|
def __init__(self, network):
|
|
super(NetWithLoss, self).__init__()
|
|
self.loss = VirtualLoss()
|
|
self.network = network
|
|
|
|
def construct(self, x, y, b):
|
|
predict = self.network(x, y, b)
|
|
return self.loss(predict)
|
|
|
|
|
|
class GradWrap(nn.Cell):
|
|
def __init__(self, network):
|
|
super(GradWrap, self).__init__()
|
|
self.network = network
|
|
|
|
def construct(self, x, y, b):
|
|
return grad_all(self.network)(x, y, b)
|
|
|
|
|
|
def compile_net(net, x, y, b):
|
|
net.set_auto_parallel()
|
|
net.set_train()
|
|
_executor.compile(net, x, y, b)
|
|
|
|
|
|
def test_matmul_pow():
|
|
class Net(nn.Cell):
|
|
def __init__(self, strategy1, strategy2):
|
|
super().__init__()
|
|
self.matmul = P.MatMul().shard(strategy1)
|
|
self.pow = P.Pow().shard(strategy2)
|
|
self.matmul2 = P.MatMul().shard(strategy1)
|
|
|
|
def construct(self, x, y, b):
|
|
out = self.matmul(x, y)
|
|
out = self.pow(out, 2.0)
|
|
out = self.matmul2(out, b)
|
|
return out
|
|
|
|
context.set_auto_parallel_context(device_num=8, global_rank=0)
|
|
strategy1 = ((2, 2), (2, 2))
|
|
strategy2 = ((4, 2), ())
|
|
net = GradWrap(NetWithLoss(Net(strategy1, strategy2)))
|
|
context.set_auto_parallel_context(parallel_mode="semi_auto_parallel")
|
|
|
|
x = Tensor(np.ones([128, 32]), dtype=ms.float32)
|
|
y = Tensor(np.ones([32, 64]), dtype=ms.float32)
|
|
b = Tensor(np.ones([64, 64]), dtype=ms.float32)
|
|
compile_net(net, x, y, b)
|
|
|
|
|
|
def test_matmul_exp():
|
|
class Net(nn.Cell):
|
|
def __init__(self, strategy1, strategy2):
|
|
super().__init__()
|
|
self.matmul = P.MatMul().shard(strategy1)
|
|
self.exp = P.Exp().shard(strategy2)
|
|
self.matmul2 = P.MatMul().shard(strategy1)
|
|
|
|
def construct(self, x, y, b):
|
|
out = self.matmul(x, y)
|
|
out = self.exp(out)
|
|
out = self.matmul2(out, b)
|
|
return out
|
|
|
|
context.set_auto_parallel_context(device_num=8, global_rank=0)
|
|
strategy1 = ((2, 2), (2, 2))
|
|
strategy2 = ((4, 2),)
|
|
net = GradWrap(NetWithLoss(Net(strategy1, strategy2)))
|
|
context.set_auto_parallel_context(parallel_mode="semi_auto_parallel")
|
|
|
|
x = Tensor(np.ones([128, 32]), dtype=ms.float32)
|
|
y = Tensor(np.ones([32, 64]), dtype=ms.float32)
|
|
b = Tensor(np.ones([64, 64]), dtype=ms.float32)
|
|
compile_net(net, x, y, b)
|
|
|
|
|
|
def test_matmul_log():
|
|
class Net(nn.Cell):
|
|
def __init__(self, strategy1, strategy2):
|
|
super().__init__()
|
|
self.matmul = P.MatMul().shard(strategy1)
|
|
self.log = P.Log().shard(strategy2)
|
|
self.matmul2 = P.MatMul().shard(strategy1)
|
|
|
|
def construct(self, x, y, b):
|
|
out = self.matmul(x, y)
|
|
out = self.log(out)
|
|
out = self.matmul2(out, b)
|
|
return out
|
|
|
|
context.set_auto_parallel_context(device_num=8, global_rank=0)
|
|
strategy1 = ((2, 2), (2, 2))
|
|
strategy2 = ((4, 2),)
|
|
net = GradWrap(NetWithLoss(Net(strategy1, strategy2)))
|
|
context.set_auto_parallel_context(parallel_mode="semi_auto_parallel")
|
|
|
|
x = Tensor(np.ones([128, 32]), dtype=ms.float32)
|
|
y = Tensor(np.ones([32, 64]), dtype=ms.float32)
|
|
b = Tensor(np.ones([64, 64]), dtype=ms.float32)
|
|
compile_net(net, x, y, b)
|
|
|
|
def test_matmul_abs():
|
|
class Net(nn.Cell):
|
|
def __init__(self, strategy1, strategy2):
|
|
super().__init__()
|
|
self.matmul = P.MatMul().shard(strategy1)
|
|
self.abs = P.Abs().shard(strategy2)
|
|
self.matmul2 = P.MatMul().shard(strategy1)
|
|
|
|
def construct(self, x, y, b):
|
|
out = self.matmul(x, y)
|
|
out = self.abs(out)
|
|
out = self.matmul2(out, b)
|
|
return out
|
|
|
|
context.set_auto_parallel_context(device_num=8, global_rank=0)
|
|
strategy1 = ((2, 2), (2, 2))
|
|
strategy2 = ((4, 2),)
|
|
net = GradWrap(NetWithLoss(Net(strategy1, strategy2)))
|
|
context.set_auto_parallel_context(parallel_mode="semi_auto_parallel")
|
|
|
|
x = Tensor(np.random.uniform(-5, 5, size=(128, 32)), dtype=ms.float32)
|
|
y = Tensor(np.random.uniform(-5, 5, size=(32, 64)), dtype=ms.float32)
|
|
b = Tensor(np.random.uniform(-5, 5, size=(64, 64)), dtype=ms.float32)
|
|
compile_net(net, x, y, b)
|
|
|
|
def test_matmul_sign():
|
|
class Net(nn.Cell):
|
|
def __init__(self, strategy1, strategy2):
|
|
super().__init__()
|
|
self.matmul = P.MatMul().shard(strategy1)
|
|
self.sign = P.Sign().shard(strategy2)
|
|
self.matmul2 = P.MatMul().shard(strategy1)
|
|
|
|
def construct(self, x, y, b):
|
|
out = self.matmul(x, y)
|
|
out = self.sign(out)
|
|
out = self.matmul2(out, b)
|
|
return out
|
|
|
|
context.set_auto_parallel_context(device_num=8, global_rank=0)
|
|
strategy1 = ((2, 2), (2, 2))
|
|
strategy2 = ((4, 2),)
|
|
net = GradWrap(NetWithLoss(Net(strategy1, strategy2)))
|
|
context.set_auto_parallel_context(parallel_mode="semi_auto_parallel")
|
|
|
|
x = Tensor(np.random.uniform(-5, 5, size=(128, 32)), dtype=ms.float32)
|
|
y = Tensor(np.random.uniform(-5, 5, size=(32, 64)), dtype=ms.float32)
|
|
b = Tensor(np.random.uniform(-5, 5, size=(64, 64)), dtype=ms.float32)
|
|
compile_net(net, x, y, b)
|
|
|
|
|
|
def test_matmul_floor():
|
|
class Net(nn.Cell):
|
|
def __init__(self, strategy1, strategy2):
|
|
super().__init__()
|
|
self.matmul = P.MatMul().shard(strategy1)
|
|
self.floor = P.Floor().shard(strategy2)
|
|
self.matmul2 = P.MatMul().shard(strategy1)
|
|
|
|
def construct(self, x, y, b):
|
|
out = self.matmul(x, y)
|
|
out = self.floor(out)
|
|
out = self.matmul2(out, b)
|
|
return out
|
|
|
|
context.set_auto_parallel_context(device_num=8, global_rank=0)
|
|
strategy1 = ((2, 2), (2, 2))
|
|
strategy2 = ((4, 2),)
|
|
net = GradWrap(NetWithLoss(Net(strategy1, strategy2)))
|
|
context.set_auto_parallel_context(parallel_mode="semi_auto_parallel")
|
|
|
|
x = Tensor(np.random.uniform(-5, 5, size=(128, 32)), dtype=ms.float32)
|
|
y = Tensor(np.random.uniform(-5, 5, size=(32, 64)), dtype=ms.float32)
|
|
b = Tensor(np.random.uniform(-5, 5, size=(64, 64)), dtype=ms.float32)
|
|
compile_net(net, x, y, b)
|
|
|
|
def test_matmul_round():
|
|
class Net(nn.Cell):
|
|
def __init__(self, strategy1, strategy2):
|
|
super().__init__()
|
|
self.matmul = P.MatMul().shard(strategy1)
|
|
self.round = P.Round().shard(strategy2)
|
|
self.matmul2 = P.MatMul().shard(strategy1)
|
|
|
|
def construct(self, x, y, b):
|
|
out = self.matmul(x, y)
|
|
out = self.round(out)
|
|
out = self.matmul2(out, b)
|
|
return out
|
|
|
|
context.set_auto_parallel_context(device_num=8, global_rank=0)
|
|
strategy1 = ((2, 2), (2, 2))
|
|
strategy2 = ((4, 2),)
|
|
net = GradWrap(NetWithLoss(Net(strategy1, strategy2)))
|
|
context.set_auto_parallel_context(parallel_mode="semi_auto_parallel")
|
|
|
|
x = Tensor(np.random.uniform(-5, 5, size=(128, 32)), dtype=ms.float32)
|
|
y = Tensor(np.random.uniform(-5, 5, size=(32, 64)), dtype=ms.float32)
|
|
b = Tensor(np.random.uniform(-5, 5, size=(64, 64)), dtype=ms.float32)
|
|
compile_net(net, x, y, b)
|
|
|
|
|
|
def test_matmul_reciprocal():
|
|
class Net(nn.Cell):
|
|
def __init__(self, strategy1, strategy2):
|
|
super().__init__()
|
|
self.matmul = P.MatMul().shard(strategy1)
|
|
self.reciprocal = P.Reciprocal().shard(strategy2)
|
|
self.matmul2 = P.MatMul().shard(strategy1)
|
|
|
|
def construct(self, x, y, b):
|
|
out = self.matmul(x, y)
|
|
out = self.reciprocal(out)
|
|
out = self.matmul2(out, b)
|
|
return out
|
|
|
|
context.set_auto_parallel_context(device_num=8, global_rank=0)
|
|
strategy1 = ((2, 2), (2, 2))
|
|
strategy2 = ((4, 2),)
|
|
net = GradWrap(NetWithLoss(Net(strategy1, strategy2)))
|
|
context.set_auto_parallel_context(parallel_mode="semi_auto_parallel")
|
|
|
|
x = Tensor(np.random.uniform(-5, 5, size=(128, 32)), dtype=ms.float32)
|
|
y = Tensor(np.random.uniform(-5, 5, size=(32, 64)), dtype=ms.float32)
|
|
b = Tensor(np.random.uniform(-5, 5, size=(64, 64)), dtype=ms.float32)
|
|
compile_net(net, x, y, b)
|
|
|
|
|
|
def test_matmul_inv():
|
|
class Net(nn.Cell):
|
|
def __init__(self, strategy1, strategy2):
|
|
super().__init__()
|
|
self.matmul = P.MatMul().shard(strategy1)
|
|
self.inv = P.Inv().shard(strategy2)
|
|
self.matmul2 = P.MatMul().shard(strategy1)
|
|
|
|
def construct(self, x, y, b):
|
|
out = self.matmul(x, y)
|
|
out = self.inv(out)
|
|
out = self.matmul2(out, b)
|
|
return out
|
|
|
|
context.set_auto_parallel_context(device_num=8, global_rank=0)
|
|
strategy1 = ((2, 2), (2, 2))
|
|
strategy2 = ((4, 2),)
|
|
net = GradWrap(NetWithLoss(Net(strategy1, strategy2)))
|
|
context.set_auto_parallel_context(parallel_mode="semi_auto_parallel")
|
|
|
|
x = Tensor(np.random.uniform(-5, 5, size=(128, 32)), dtype=ms.float32)
|
|
y = Tensor(np.random.uniform(-5, 5, size=(32, 64)), dtype=ms.float32)
|
|
b = Tensor(np.random.uniform(-5, 5, size=(64, 64)), dtype=ms.float32)
|
|
compile_net(net, x, y, b)
|
|
|
|
|
|
def test_matmul_rsqrt():
|
|
class Net(nn.Cell):
|
|
def __init__(self, strategy1, strategy2):
|
|
super().__init__()
|
|
self.matmul = P.MatMul().shard(strategy1)
|
|
self.rsqrt = P.Rsqrt().shard(strategy2)
|
|
self.matmul2 = P.MatMul().shard(strategy1)
|
|
|
|
def construct(self, x, y, b):
|
|
out = self.matmul(x, y)
|
|
out = self.rsqrt(out)
|
|
out = self.matmul2(out, b)
|
|
return out
|
|
|
|
context.set_auto_parallel_context(device_num=8, global_rank=0)
|
|
strategy1 = ((2, 2), (2, 2))
|
|
strategy2 = ((4, 2),)
|
|
net = GradWrap(NetWithLoss(Net(strategy1, strategy2)))
|
|
context.set_auto_parallel_context(parallel_mode="semi_auto_parallel")
|
|
|
|
x = Tensor(np.random.uniform(-5, 5, size=(128, 32)), dtype=ms.float32)
|
|
y = Tensor(np.random.uniform(-5, 5, size=(32, 64)), dtype=ms.float32)
|
|
b = Tensor(np.random.uniform(-5, 5, size=(64, 64)), dtype=ms.float32)
|
|
compile_net(net, x, y, b)
|
|
|
|
|
|
def test_matmul_tan():
|
|
class Net(nn.Cell):
|
|
def __init__(self, strategy1, strategy2):
|
|
super().__init__()
|
|
self.matmul = P.MatMul().shard(strategy1)
|
|
self.tan = P.Tan().shard(strategy2)
|
|
self.matmul2 = P.MatMul().shard(strategy1)
|
|
|
|
def construct(self, x, y, b):
|
|
out = self.matmul(x, y)
|
|
out = self.tan(out)
|
|
out = self.matmul2(out, b)
|
|
return out
|
|
|
|
context.set_auto_parallel_context(device_num=8, global_rank=0)
|
|
strategy1 = ((2, 2), (2, 2))
|
|
strategy2 = ((4, 2),)
|
|
net = GradWrap(NetWithLoss(Net(strategy1, strategy2)))
|
|
context.set_auto_parallel_context(parallel_mode="semi_auto_parallel")
|
|
|
|
x = Tensor(np.random.uniform(-5, 5, size=(128, 32)), dtype=ms.float32)
|
|
y = Tensor(np.random.uniform(-5, 5, size=(32, 64)), dtype=ms.float32)
|
|
b = Tensor(np.random.uniform(-5, 5, size=(64, 64)), dtype=ms.float32)
|
|
compile_net(net, x, y, b)
|
|
|
|
|
|
def test_matmul_sin():
|
|
class Net(nn.Cell):
|
|
def __init__(self, strategy1, strategy2):
|
|
super().__init__()
|
|
self.matmul = P.MatMul().shard(strategy1)
|
|
self.sin = P.Sin().shard(strategy2)
|
|
self.matmul2 = P.MatMul().shard(strategy1)
|
|
|
|
def construct(self, x, y, b):
|
|
out = self.matmul(x, y)
|
|
out = self.sin(out)
|
|
out = self.matmul2(out, b)
|
|
return out
|
|
|
|
context.set_auto_parallel_context(device_num=8, global_rank=0)
|
|
strategy1 = ((2, 2), (2, 2))
|
|
strategy2 = ((4, 2),)
|
|
net = GradWrap(NetWithLoss(Net(strategy1, strategy2)))
|
|
context.set_auto_parallel_context(parallel_mode="semi_auto_parallel")
|
|
|
|
x = Tensor(np.random.uniform(-5, 5, size=(128, 32)), dtype=ms.float32)
|
|
y = Tensor(np.random.uniform(-5, 5, size=(32, 64)), dtype=ms.float32)
|
|
b = Tensor(np.random.uniform(-5, 5, size=(64, 64)), dtype=ms.float32)
|
|
compile_net(net, x, y, b)
|
|
|
|
|
|
def test_matmul_sinh():
|
|
class Net(nn.Cell):
|
|
def __init__(self, strategy1, strategy2):
|
|
super().__init__()
|
|
self.matmul = P.MatMul().shard(strategy1)
|
|
self.sinh = P.Sinh().shard(strategy2)
|
|
self.matmul2 = P.MatMul().shard(strategy1)
|
|
|
|
def construct(self, x, y, b):
|
|
out = self.matmul(x, y)
|
|
out = self.sinh(out)
|
|
out = self.matmul2(out, b)
|
|
return out
|
|
|
|
context.set_auto_parallel_context(device_num=8, global_rank=0)
|
|
strategy1 = ((2, 2), (2, 2))
|
|
strategy2 = ((4, 2),)
|
|
net = GradWrap(NetWithLoss(Net(strategy1, strategy2)))
|
|
context.set_auto_parallel_context(parallel_mode="semi_auto_parallel")
|
|
|
|
x = Tensor(np.random.uniform(-5, 5, size=(128, 32)), dtype=ms.float32)
|
|
y = Tensor(np.random.uniform(-5, 5, size=(32, 64)), dtype=ms.float32)
|
|
b = Tensor(np.random.uniform(-5, 5, size=(64, 64)), dtype=ms.float32)
|
|
compile_net(net, x, y, b)
|
|
|
|
|
|
def test_matmul_log1p():
|
|
class Net(nn.Cell):
|
|
def __init__(self, strategy1, strategy2):
|
|
super().__init__()
|
|
self.matmul = P.MatMul().shard(strategy1)
|
|
self.log1p = P.Log1p().shard(strategy2)
|
|
self.matmul2 = P.MatMul().shard(strategy1)
|
|
|
|
def construct(self, x, y, b):
|
|
out = self.matmul(x, y)
|
|
out = self.log1p(out)
|
|
out = self.matmul2(out, b)
|
|
return out
|
|
|
|
context.set_auto_parallel_context(device_num=8, global_rank=0)
|
|
strategy1 = ((2, 2), (2, 2))
|
|
strategy2 = ((4, 2),)
|
|
net = GradWrap(NetWithLoss(Net(strategy1, strategy2)))
|
|
context.set_auto_parallel_context(parallel_mode="semi_auto_parallel")
|
|
|
|
x = Tensor(np.random.uniform(-5, 5, size=(128, 32)), dtype=ms.float32)
|
|
y = Tensor(np.random.uniform(-5, 5, size=(32, 64)), dtype=ms.float32)
|
|
b = Tensor(np.random.uniform(-5, 5, size=(64, 64)), dtype=ms.float32)
|
|
compile_net(net, x, y, b)
|
|
|
|
|
|
def test_matmul_expm1():
|
|
class Net(nn.Cell):
|
|
def __init__(self, strategy1, strategy2):
|
|
super().__init__()
|
|
self.matmul = P.MatMul().shard(strategy1)
|
|
self.expm1 = P.Expm1().shard(strategy2)
|
|
self.matmul2 = P.MatMul().shard(strategy1)
|
|
|
|
def construct(self, x, y, b):
|
|
out = self.matmul(x, y)
|
|
out = self.expm1(out)
|
|
out = self.matmul2(out, b)
|
|
return out
|
|
|
|
context.set_auto_parallel_context(device_num=8, global_rank=0)
|
|
strategy1 = ((2, 2), (2, 2))
|
|
strategy2 = ((4, 2),)
|
|
net = GradWrap(NetWithLoss(Net(strategy1, strategy2)))
|
|
context.set_auto_parallel_context(parallel_mode="semi_auto_parallel")
|
|
|
|
x = Tensor(np.random.uniform(-5, 5, size=(128, 32)), dtype=ms.float32)
|
|
y = Tensor(np.random.uniform(-5, 5, size=(32, 64)), dtype=ms.float32)
|
|
b = Tensor(np.random.uniform(-5, 5, size=(64, 64)), dtype=ms.float32)
|
|
compile_net(net, x, y, b)
|
|
|
|
|
|
def test_matmul_cosh():
|
|
class Net(nn.Cell):
|
|
def __init__(self, strategy1, strategy2):
|
|
super().__init__()
|
|
self.matmul = P.MatMul().shard(strategy1)
|
|
self.cosh = P.Cosh().shard(strategy2)
|
|
self.matmul2 = P.MatMul().shard(strategy1)
|
|
|
|
def construct(self, x, y, b):
|
|
out = self.matmul(x, y)
|
|
out = self.cosh(out)
|
|
out = self.matmul2(out, b)
|
|
return out
|
|
|
|
context.set_auto_parallel_context(device_num=8, global_rank=0)
|
|
strategy1 = ((2, 2), (2, 2))
|
|
strategy2 = ((4, 2),)
|
|
net = GradWrap(NetWithLoss(Net(strategy1, strategy2)))
|
|
context.set_auto_parallel_context(parallel_mode="semi_auto_parallel")
|
|
|
|
x = Tensor(np.random.uniform(-5, 5, size=(128, 32)), dtype=ms.float32)
|
|
y = Tensor(np.random.uniform(-5, 5, size=(32, 64)), dtype=ms.float32)
|
|
b = Tensor(np.random.uniform(-5, 5, size=(64, 64)), dtype=ms.float32)
|
|
compile_net(net, x, y, b)
|
|
|
|
def test_matmul_erf():
|
|
class Net(nn.Cell):
|
|
def __init__(self, strategy1, strategy2):
|
|
super().__init__()
|
|
self.matmul = P.MatMul().shard(strategy1)
|
|
self.erf = P.Erf().shard(strategy2)
|
|
self.matmul2 = P.MatMul().shard(strategy1)
|
|
|
|
def construct(self, x, y, b):
|
|
out = self.matmul(x, y)
|
|
out = self.erf(out)
|
|
out = self.matmul2(out, b)
|
|
return out
|
|
|
|
context.set_auto_parallel_context(device_num=8, global_rank=0)
|
|
strategy1 = ((2, 2), (2, 2))
|
|
strategy2 = ((4, 2),)
|
|
net = GradWrap(NetWithLoss(Net(strategy1, strategy2)))
|
|
context.set_auto_parallel_context(parallel_mode="semi_auto_parallel")
|
|
|
|
x = Tensor(np.random.uniform(1, 5, size=(128, 32)), dtype=ms.float32)
|
|
y = Tensor(np.random.uniform(1, 5, size=(32, 64)), dtype=ms.float32)
|
|
b = Tensor(np.random.uniform(1, 5, size=(64, 64)), dtype=ms.float32)
|
|
compile_net(net, x, y, b)
|
|
|
|
|
|
def test_matmul_erfc():
|
|
class Net(nn.Cell):
|
|
def __init__(self, strategy1, strategy2):
|
|
super().__init__()
|
|
self.matmul = P.MatMul().shard(strategy1)
|
|
self.erfc = P.Erfc().shard(strategy2)
|
|
self.matmul2 = P.MatMul().shard(strategy1)
|
|
|
|
def construct(self, x, y, b):
|
|
out = self.matmul(x, y)
|
|
out = self.erfc(out)
|
|
out = self.matmul2(out, b)
|
|
return out
|
|
|
|
context.set_auto_parallel_context(device_num=8, global_rank=0)
|
|
strategy1 = ((2, 2), (2, 2))
|
|
strategy2 = ((4, 2),)
|
|
net = GradWrap(NetWithLoss(Net(strategy1, strategy2)))
|
|
context.set_auto_parallel_context(parallel_mode="semi_auto_parallel")
|
|
|
|
x = Tensor(np.random.uniform(1, 5, size=(128, 32)), dtype=ms.float32)
|
|
y = Tensor(np.random.uniform(1, 5, size=(32, 64)), dtype=ms.float32)
|
|
b = Tensor(np.random.uniform(1, 5, size=(64, 64)), dtype=ms.float32)
|
|
compile_net(net, x, y, b)
|
|
|
|
|
|
def test_matmul_zeroslike():
|
|
class Net(nn.Cell):
|
|
def __init__(self, strategy1, strategy2):
|
|
super().__init__()
|
|
self.matmul = P.MatMul().shard(strategy1)
|
|
self.zeroslike = P.ZerosLike().shard(strategy2)
|
|
self.matmul2 = P.MatMul().shard(strategy1)
|
|
|
|
def construct(self, x, y, b):
|
|
out = self.matmul(x, y)
|
|
out = self.zeroslike(out)
|
|
out = self.matmul2(out, b)
|
|
return out
|
|
|
|
context.set_auto_parallel_context(device_num=8, global_rank=0)
|
|
strategy1 = ((2, 2), (2, 2))
|
|
strategy2 = ((4, 2),)
|
|
net = GradWrap(NetWithLoss(Net(strategy1, strategy2)))
|
|
context.set_auto_parallel_context(parallel_mode="semi_auto_parallel")
|
|
|
|
x = Tensor(np.random.uniform(1, 5, size=(128, 32)), dtype=ms.float32)
|
|
y = Tensor(np.random.uniform(1, 5, size=(32, 64)), dtype=ms.float32)
|
|
b = Tensor(np.random.uniform(1, 5, size=(64, 64)), dtype=ms.float32)
|
|
compile_net(net, x, y, b)
|
|
|
|
|
|
def test_matmul_oneslike():
|
|
class Net(nn.Cell):
|
|
def __init__(self, strategy1, strategy2):
|
|
super().__init__()
|
|
self.matmul = P.MatMul().shard(strategy1)
|
|
self.oneslike = P.OnesLike().shard(strategy2)
|
|
self.matmul2 = P.MatMul().shard(strategy1)
|
|
|
|
def construct(self, x, y, b):
|
|
out = self.matmul(x, y)
|
|
out = self.oneslike(out)
|
|
out = self.matmul2(out, b)
|
|
return out
|
|
|
|
context.set_auto_parallel_context(device_num=8, global_rank=0)
|
|
strategy1 = ((2, 2), (2, 2))
|
|
strategy2 = ((4, 2),)
|
|
net = GradWrap(NetWithLoss(Net(strategy1, strategy2)))
|
|
context.set_auto_parallel_context(parallel_mode="semi_auto_parallel")
|
|
|
|
x = Tensor(np.random.uniform(1, 5, size=(128, 32)), dtype=ms.float32)
|
|
y = Tensor(np.random.uniform(1, 5, size=(32, 64)), dtype=ms.float32)
|
|
b = Tensor(np.random.uniform(1, 5, size=(64, 64)), dtype=ms.float32)
|
|
compile_net(net, x, y, b)
|
|
|
|
|
|
def test_matmul_BesselI0e():
|
|
class Net(nn.Cell):
|
|
def __init__(self, strategy1, strategy2):
|
|
super().__init__()
|
|
self.matmul = P.MatMul().shard(strategy1)
|
|
self.BesselI0e = P.BesselI0e().shard(strategy2)
|
|
self.matmul2 = P.MatMul().shard(strategy1)
|
|
|
|
def construct(self, x, y, b):
|
|
out = self.matmul(x, y)
|
|
out = self.BesselI0e(out)
|
|
out = self.matmul2(out, b)
|
|
return out
|
|
|
|
context.set_auto_parallel_context(device_num=8, global_rank=0)
|
|
strategy1 = ((2, 2), (2, 2))
|
|
strategy2 = ((4, 2),)
|
|
net = GradWrap(NetWithLoss(Net(strategy1, strategy2)))
|
|
context.set_auto_parallel_context(parallel_mode="semi_auto_parallel")
|
|
|
|
x = Tensor(np.random.uniform(1, 5, size=(128, 32)), dtype=ms.float32)
|
|
y = Tensor(np.random.uniform(1, 5, size=(32, 64)), dtype=ms.float32)
|
|
b = Tensor(np.random.uniform(1, 5, size=(64, 64)), dtype=ms.float32)
|
|
compile_net(net, x, y, b)
|
|
|
|
|
|
def test_matmul_BesselI1e():
|
|
class Net(nn.Cell):
|
|
def __init__(self, strategy1, strategy2):
|
|
super().__init__()
|
|
self.matmul = P.MatMul().shard(strategy1)
|
|
self.BesselI1e = P.BesselI1e().shard(strategy2)
|
|
self.matmul2 = P.MatMul().shard(strategy1)
|
|
|
|
def construct(self, x, y, b):
|
|
out = self.matmul(x, y)
|
|
out = self.BesselI1e(out)
|
|
out = self.matmul2(out, b)
|
|
return out
|
|
|
|
context.set_auto_parallel_context(device_num=8, global_rank=0)
|
|
strategy1 = ((2, 2), (2, 2))
|
|
strategy2 = ((4, 2),)
|
|
net = GradWrap(NetWithLoss(Net(strategy1, strategy2)))
|
|
context.set_auto_parallel_context(parallel_mode="semi_auto_parallel")
|
|
|
|
x = Tensor(np.random.uniform(1, 5, size=(128, 32)), dtype=ms.float32)
|
|
y = Tensor(np.random.uniform(1, 5, size=(32, 64)), dtype=ms.float32)
|
|
b = Tensor(np.random.uniform(1, 5, size=(64, 64)), dtype=ms.float32)
|
|
compile_net(net, x, y, b)
|
|
|
|
|
|
def test_matmul_ceil():
|
|
class Net(nn.Cell):
|
|
def __init__(self, strategy1, strategy2):
|
|
super().__init__()
|
|
self.matmul = P.MatMul().shard(strategy1)
|
|
self.Ceil = P.Ceil().shard(strategy2)
|
|
self.matmul2 = P.MatMul().shard(strategy1)
|
|
|
|
def construct(self, x, y, b):
|
|
out = self.matmul(x, y)
|
|
out = self.Ceil(out)
|
|
out = self.matmul2(out, b)
|
|
return out
|
|
|
|
context.set_auto_parallel_context(device_num=8, global_rank=0)
|
|
strategy1 = ((2, 2), (2, 2))
|
|
strategy2 = ((4, 2),)
|
|
net = GradWrap(NetWithLoss(Net(strategy1, strategy2)))
|
|
context.set_auto_parallel_context(parallel_mode="semi_auto_parallel")
|
|
|
|
x = Tensor(np.random.uniform(-5, 5, size=(128, 32)), dtype=ms.float32)
|
|
y = Tensor(np.random.uniform(-5, 5, size=(32, 64)), dtype=ms.float32)
|
|
b = Tensor(np.random.uniform(-5, 5, size=(64, 64)), dtype=ms.float32)
|
|
compile_net(net, x, y, b)
|
|
|
|
|
|
def test_matmul_atan():
|
|
class Net(nn.Cell):
|
|
def __init__(self, strategy1, strategy2):
|
|
super().__init__()
|
|
self.matmul = P.MatMul().shard(strategy1)
|
|
self.atan = P.Atan().shard(strategy2)
|
|
self.matmul2 = P.MatMul().shard(strategy1)
|
|
|
|
def construct(self, x, y, b):
|
|
out = self.matmul(x, y)
|
|
out = self.atan(out)
|
|
out = self.matmul2(out, b)
|
|
return out
|
|
|
|
context.set_auto_parallel_context(device_num=8, global_rank=0)
|
|
strategy1 = ((2, 2), (2, 2))
|
|
strategy2 = ((4, 2),)
|
|
net = GradWrap(NetWithLoss(Net(strategy1, strategy2)))
|
|
context.set_auto_parallel_context(parallel_mode="semi_auto_parallel")
|
|
|
|
x = Tensor(np.random.uniform(-5, 5, size=(128, 32)), dtype=ms.float32)
|
|
y = Tensor(np.random.uniform(-5, 5, size=(32, 64)), dtype=ms.float32)
|
|
b = Tensor(np.random.uniform(-5, 5, size=(64, 64)), dtype=ms.float32)
|
|
compile_net(net, x, y, b)
|
|
|
|
|
|
def test_matmul_Atanh():
|
|
class Net(nn.Cell):
|
|
def __init__(self, strategy1, strategy2):
|
|
super().__init__()
|
|
self.matmul = P.MatMul().shard(strategy1)
|
|
self.atanh = P.Atanh().shard(strategy2)
|
|
self.matmul2 = P.MatMul().shard(strategy1)
|
|
|
|
def construct(self, x, y, b):
|
|
out = self.matmul(x, y)
|
|
out = self.atanh(out)
|
|
out = self.matmul2(out, b)
|
|
return out
|
|
|
|
context.set_auto_parallel_context(device_num=8, global_rank=0)
|
|
strategy1 = ((2, 2), (2, 2))
|
|
strategy2 = ((4, 2),)
|
|
net = GradWrap(NetWithLoss(Net(strategy1, strategy2)))
|
|
context.set_auto_parallel_context(parallel_mode="semi_auto_parallel")
|
|
|
|
x = Tensor(np.random.uniform(-5, 5, size=(128, 32)), dtype=ms.float32)
|
|
y = Tensor(np.random.uniform(-5, 5, size=(32, 64)), dtype=ms.float32)
|
|
b = Tensor(np.random.uniform(-5, 5, size=(64, 64)), dtype=ms.float32)
|
|
compile_net(net, x, y, b)
|
|
|
|
|
|
def test_matmul_asin():
|
|
class Net(nn.Cell):
|
|
def __init__(self, strategy1, strategy2):
|
|
super().__init__()
|
|
self.matmul = P.MatMul().shard(strategy1)
|
|
self.asin = P.Asin().shard(strategy2)
|
|
self.matmul2 = P.MatMul().shard(strategy1)
|
|
|
|
def construct(self, x, y, b):
|
|
out = self.matmul(x, y)
|
|
out = self.asin(out)
|
|
out = self.matmul2(out, b)
|
|
return out
|
|
|
|
context.set_auto_parallel_context(device_num=8, global_rank=0)
|
|
strategy1 = ((2, 2), (2, 2))
|
|
strategy2 = ((4, 2),)
|
|
net = GradWrap(NetWithLoss(Net(strategy1, strategy2)))
|
|
context.set_auto_parallel_context(parallel_mode="semi_auto_parallel")
|
|
|
|
x = Tensor(np.random.uniform(-5, 5, size=(128, 32)), dtype=ms.float32)
|
|
y = Tensor(np.random.uniform(-5, 5, size=(32, 64)), dtype=ms.float32)
|
|
b = Tensor(np.random.uniform(-5, 5, size=(64, 64)), dtype=ms.float32)
|
|
compile_net(net, x, y, b)
|
|
|
|
|
|
def test_matmul_asinh():
|
|
class Net(nn.Cell):
|
|
def __init__(self, strategy1, strategy2):
|
|
super().__init__()
|
|
self.matmul = P.MatMul().shard(strategy1)
|
|
self.asinh = P.Asinh().shard(strategy2)
|
|
self.matmul2 = P.MatMul().shard(strategy1)
|
|
|
|
def construct(self, x, y, b):
|
|
out = self.matmul(x, y)
|
|
out = self.asinh(out)
|
|
out = self.matmul2(out, b)
|
|
return out
|
|
|
|
context.set_auto_parallel_context(device_num=8, global_rank=0)
|
|
strategy1 = ((2, 2), (2, 2))
|
|
strategy2 = ((4, 2),)
|
|
net = GradWrap(NetWithLoss(Net(strategy1, strategy2)))
|
|
context.set_auto_parallel_context(parallel_mode="semi_auto_parallel")
|
|
|
|
x = Tensor(np.random.uniform(-5, 5, size=(128, 32)), dtype=ms.float32)
|
|
y = Tensor(np.random.uniform(-5, 5, size=(32, 64)), dtype=ms.float32)
|
|
b = Tensor(np.random.uniform(-5, 5, size=(64, 64)), dtype=ms.float32)
|
|
compile_net(net, x, y, b)
|
|
|
|
|
|
def test_matmul_acosh():
|
|
class Net(nn.Cell):
|
|
def __init__(self, strategy1, strategy2):
|
|
super().__init__()
|
|
self.matmul = P.MatMul().shard(strategy1)
|
|
self.acosh = P.Acosh().shard(strategy2)
|
|
self.matmul2 = P.MatMul().shard(strategy1)
|
|
|
|
def construct(self, x, y, b):
|
|
out = self.matmul(x, y)
|
|
out = self.acosh(out)
|
|
out = self.matmul2(out, b)
|
|
return out
|
|
|
|
context.set_auto_parallel_context(device_num=8, global_rank=0)
|
|
strategy1 = ((2, 2), (2, 2))
|
|
strategy2 = ((4, 2),)
|
|
net = GradWrap(NetWithLoss(Net(strategy1, strategy2)))
|
|
context.set_auto_parallel_context(parallel_mode="semi_auto_parallel")
|
|
|
|
x = Tensor(np.random.uniform(-5, 5, size=(128, 32)), dtype=ms.float32)
|
|
y = Tensor(np.random.uniform(-5, 5, size=(32, 64)), dtype=ms.float32)
|
|
b = Tensor(np.random.uniform(-5, 5, size=(64, 64)), dtype=ms.float32)
|
|
compile_net(net, x, y, b)
|
|
|
|
|
|
def test_matmul_logical_not():
|
|
class Net(nn.Cell):
|
|
def __init__(self, strategy1, strategy2, strategy3):
|
|
super().__init__()
|
|
self.matmul = P.MatMul().shard(strategy1)
|
|
self.logicalnot = P.LogicalNot().shard(strategy2)
|
|
self.equal = P.Equal().shard(strategy3)
|
|
|
|
def construct(self, x, y, b):
|
|
out = self.matmul(x, y)
|
|
out = self.equal(out, b)
|
|
out = self.logicalnot(out)
|
|
return out
|
|
|
|
context.set_auto_parallel_context(device_num=8, global_rank=0)
|
|
strategy1 = ((2, 2), (2, 2))
|
|
strategy2 = ((4, 2),)
|
|
strategy3 = ((4, 2), (4, 2))
|
|
net = GradWrap(NetWithLoss(Net(strategy1, strategy2, strategy3)))
|
|
context.set_auto_parallel_context(parallel_mode="semi_auto_parallel")
|
|
|
|
x = Tensor(np.ones([128, 32]), dtype=ms.float32)
|
|
y = Tensor(np.ones([32, 64]), dtype=ms.float32)
|
|
b = Tensor(np.ones([128, 64]), dtype=ms.float32)
|
|
compile_net(net, x, y, b)
|
|
|
|
|
|
def test_matmul_cast():
|
|
class Net(nn.Cell):
|
|
def __init__(self, strategy1, strategy2, strategy3):
|
|
super().__init__()
|
|
self.matmul = P.MatMul().shard(strategy1)
|
|
self.cast = P.Cast().shard(strategy2)
|
|
self.matmul2 = P.MatMul().shard(strategy3)
|
|
|
|
def construct(self, x, y, b):
|
|
out = self.matmul(x, y)
|
|
b = self.cast(b, ms.float32)
|
|
out = self.matmul2(out, b)
|
|
return out
|
|
|
|
context.set_auto_parallel_context(device_num=8, global_rank=0)
|
|
strategy1 = ((2, 2), (2, 2))
|
|
strategy2 = ((4, 2),)
|
|
strategy3 = ((1, 4), (4, 2))
|
|
net = GradWrap(NetWithLoss(Net(strategy1, strategy2, strategy3)))
|
|
context.set_auto_parallel_context(parallel_mode="semi_auto_parallel")
|
|
|
|
x = Tensor(np.ones([128, 32]), dtype=ms.float32)
|
|
y = Tensor(np.ones([32, 64]), dtype=ms.float32)
|
|
b = Tensor(np.ones([64, 64]), dtype=ms.int32)
|
|
compile_net(net, x, y, b)
|
|
|
|
|
|
def test_gradient_fp32_sync():
|
|
class Net(nn.Cell):
|
|
def __init__(self, strategy1):
|
|
super().__init__()
|
|
self.matmul = P.MatMul().shard(strategy1)
|
|
self.cast = P.Cast()
|
|
|
|
def construct(self, x, y, b):
|
|
out = self.matmul(x, y)
|
|
b = self.cast(b, ms.float32)
|
|
out = self.matmul(out, b)
|
|
return out
|
|
|
|
context.set_auto_parallel_context(device_num=8, global_rank=0, gradient_fp32_sync=True)
|
|
strategy1 = ((2, 2), (2, 2))
|
|
net = GradWrap(NetWithLoss(Net(strategy1)))
|
|
context.set_auto_parallel_context(parallel_mode="semi_auto_parallel")
|
|
|
|
x = Tensor(np.ones([128, 32]), dtype=ms.float32)
|
|
y = Tensor(np.ones([32, 64]), dtype=ms.float32)
|
|
b = Tensor(np.ones([64, 64]), dtype=ms.float16)
|
|
compile_net(net, x, y, b)
|
|
|
|
|
|
def test_gradient_fp32_sync1():
|
|
class Net(nn.Cell):
|
|
def __init__(self, strategy1):
|
|
super().__init__()
|
|
self.matmul = P.MatMul().shard(strategy1)
|
|
self.cast = P.Cast()
|
|
|
|
def construct(self, x, y, b):
|
|
out = self.matmul(x, y)
|
|
b = self.cast(b, ms.float16)
|
|
out = self.matmul(out, b)
|
|
return out
|
|
|
|
context.set_auto_parallel_context(device_num=8, global_rank=0, gradient_fp32_sync=True)
|
|
strategy1 = ((2, 2), (2, 2))
|
|
net = GradWrap(NetWithLoss(Net(strategy1)))
|
|
context.set_auto_parallel_context(parallel_mode="semi_auto_parallel")
|
|
|
|
x = Tensor(np.ones([128, 32]), dtype=ms.float16)
|
|
y = Tensor(np.ones([32, 64]), dtype=ms.float16)
|
|
b = Tensor(np.ones([64, 64]), dtype=ms.float32)
|
|
compile_net(net, x, y, b)
|
|
|
|
|
|
def test_gradient_fp32_sync2():
|
|
class Net(nn.Cell):
|
|
def __init__(self, strategy1):
|
|
super().__init__()
|
|
self.matmul = P.MatMul().shard(strategy1)
|
|
self.cast = P.Cast()
|
|
|
|
def construct(self, x, y, b):
|
|
out = self.matmul(x, y)
|
|
b = self.cast(b, ms.float16)
|
|
out = self.matmul(out, b)
|
|
return out
|
|
|
|
context.set_auto_parallel_context(device_num=8, global_rank=0, gradient_fp32_sync=False)
|
|
strategy1 = ((2, 2), (2, 2))
|
|
net = GradWrap(NetWithLoss(Net(strategy1)))
|
|
context.set_auto_parallel_context(parallel_mode="semi_auto_parallel")
|
|
|
|
x = Tensor(np.ones([128, 32]), dtype=ms.float16)
|
|
y = Tensor(np.ones([32, 64]), dtype=ms.float16)
|
|
b = Tensor(np.ones([64, 64]), dtype=ms.float32)
|
|
compile_net(net, x, y, b)
|
|
|
|
|
|
def test_gradient_fp32_sync3():
|
|
class Net(nn.Cell):
|
|
def __init__(self, strategy1):
|
|
super().__init__()
|
|
self.matmul = P.MatMul().shard(strategy1)
|
|
self.cast = P.Cast()
|
|
|
|
def construct(self, x, y, b):
|
|
out = self.matmul(x, y)
|
|
b = self.cast(b, ms.float16)
|
|
out = self.matmul(out, b)
|
|
return out
|
|
|
|
context.set_auto_parallel_context(device_num=8, global_rank=0)
|
|
strategy1 = ((2, 2), (2, 2))
|
|
net = GradWrap(NetWithLoss(Net(strategy1)))
|
|
context.set_auto_parallel_context(parallel_mode="semi_auto_parallel")
|
|
|
|
x = Tensor(np.ones([128, 32]), dtype=ms.float16)
|
|
y = Tensor(np.ones([32, 64]), dtype=ms.float16)
|
|
b = Tensor(np.ones([64, 64]), dtype=ms.float32)
|
|
compile_net(net, x, y, b)
|
|
|
|
|
|
def test_mul_two_cast():
|
|
class Net(nn.Cell):
|
|
def __init__(self, strategy1, strategy2, strategy3):
|
|
super().__init__()
|
|
self.mul = P.Mul().shard(strategy1)
|
|
self.mul2 = P.Mul().shard(strategy2)
|
|
self.cast = P.Cast().shard(strategy3)
|
|
self.cast2 = P.Cast().shard(strategy3)
|
|
|
|
def construct(self, x, y, b):
|
|
out = self.mul(x, y)
|
|
out = self.mul2(out, b)
|
|
out = self.cast(out, ms.int32)
|
|
out = self.cast2(out, ms.bool_)
|
|
return out
|
|
|
|
context.set_auto_parallel_context(device_num=8, global_rank=0)
|
|
strategy1 = ((2, 2), (2, 2))
|
|
strategy2 = ((8, 1), (8, 1))
|
|
strategy3 = ((8, 1),)
|
|
net = GradWrap(Net(strategy1, strategy2, strategy3))
|
|
context.set_auto_parallel_context(parallel_mode="semi_auto_parallel")
|
|
|
|
x = Tensor(np.ones([128, 32]), dtype=ms.float32)
|
|
y = Tensor(np.ones([128, 32]), dtype=ms.float32)
|
|
b = Tensor(np.ones([128, 32]), dtype=ms.float32)
|
|
compile_net(net, x, y, b)
|