from train_utils import save_inout, train_wrap
from official.cv.xception.src.Xception import Xception
import mindspore.common.dtype as mstype
from mindspore import context, Tensor, nn
from mindspore.train.serialization import export
context.set_context(mode=context.PYNATIVE_MODE,
device_target="GPU",
save_graphs=False)
n = Xception(num_classes=1000)
n.dropout = nn.Dropout(keep_prob=1.0)
loss_fn = nn.SoftmaxCrossEntropyWithLogits(sparse=False)
optimizer = nn.SGD(n.trainable_params(),
learning_rate=0.01,
momentum=0.9,
dampening=0.0,
weight_decay=0.0,
nesterov=True,
loss_scale=1.0)
net = train_wrap(n, loss_fn, optimizer)
batch = 2
x = Tensor(np.random.randn(batch, 3, 299, 299), mstype.float32)
label = Tensor(np.zeros([batch, 1000]).astype(np.float32))
export(net, x, label, file_name="mindir/xception_train", file_format='MINDIR')
if len(sys.argv) > 1:
save_inout(sys.argv[1] + "xception", x, label, n, net)
# See the License for the specific language governing permissions and
# limitations under the License.
# ============================================================================
"""vgg_train_export."""
import sys
import numpy as np
from train_utils import save_inout, train_wrap
from official.cv.vgg16.src.vgg import vgg16
import mindspore.common.dtype as mstype
from mindspore import context, Tensor, nn
from mindspore.train.serialization import export
context.set_context(mode=context.PYNATIVE_MODE,
device_target="GPU",
save_graphs=False)
batch = 2
n = vgg16(num_classes=10)
loss_fn = nn.SoftmaxCrossEntropyWithLogits(sparse=False)
optimizer = nn.Momentum(n.trainable_params(), 0.01, 0.9, use_nesterov=False)
net = train_wrap(n, loss_fn, optimizer)
x = Tensor(np.random.randn(batch, 3, 224, 224), mstype.float32)
label = Tensor(np.zeros([batch, 10]).astype(np.float32))
export(net, x, label, file_name="mindir/vgg_train", file_format='MINDIR')
if len(sys.argv) > 1:
save_inout(sys.argv[1] + "vgg", x, label, n, net)
# 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.
# ============================================================================
"""effnet_train_export."""
import sys
import numpy as np
from train_utils import save_inout, train_wrap
from effnet import effnet
import mindspore.common.dtype as mstype
from mindspore import context, Tensor, nn
from mindspore.train.serialization import export
context.set_context(mode=context.PYNATIVE_MODE, device_target="GPU", save_graphs=False)
n = effnet(num_classes=10)
loss_fn = nn.SoftmaxCrossEntropyWithLogits(sparse=False)
optimizer = nn.SGD(n.trainable_params(), learning_rate=0.01, momentum=0.9, dampening=0.0, weight_decay=0.0,
nesterov=True, loss_scale=1.0)
net = train_wrap(n, loss_fn, optimizer)
x = Tensor(np.random.randn(2, 3, 224, 224), mstype.float32)
label = Tensor(np.zeros([2, 10]).astype(np.float32))
export(net, x, label, file_name="mindir/effnet_train", file_format='MINDIR')
if len(sys.argv) > 1:
save_inout(sys.argv[1] + "effnet", x, label, n, net)
import numpy as np
from train_utils import save_inout, train_wrap
from official.cv.lenet.src.lenet import LeNet5
import mindspore.common.dtype as mstype
from mindspore import context, Tensor, nn
from mindspore.train.serialization import export
context.set_context(mode=context.PYNATIVE_MODE,
device_target="GPU",
save_graphs=False)
n = LeNet5()
loss_fn = nn.MSELoss()
optimizer = nn.Adam(n.trainable_params(),
learning_rate=1e-2,
beta1=0.5,
beta2=0.7,
eps=1e-2,
use_locking=True,
use_nesterov=False,
weight_decay=0.0,
loss_scale=0.3)
net = train_wrap(n, loss_fn, optimizer)
x = Tensor(np.random.randn(32, 1, 32, 32), mstype.float32)
label = Tensor(np.zeros([32, 10]).astype(np.float32))
export(net, x, label, file_name="mindir/lenet_train", file_format='MINDIR')
if len(sys.argv) > 1:
save_inout(sys.argv[1] + "lenet", x, label, n, net, sparse=False)
device_target="GPU",
save_graphs=False)
# Mini alexnet is designed for MNIST data
batch = 2
number_of_classes = 10
n = AlexNet(phase='test')
loss_fn = nn.SoftmaxCrossEntropyWithLogits(sparse=False)
optimizer = nn.Adam(n.trainable_params(),
learning_rate=1e-3,
beta1=0.9,
beta2=0.999,
eps=1e-8,
use_locking=False,
use_nesterov=False,
weight_decay=0.0,
loss_scale=1.0)
net = train_wrap(n, loss_fn, optimizer)
x = Tensor(np.ones([batch, 1, 32, 32]).astype(np.float32) * 0.01)
label = Tensor(np.zeros([batch, number_of_classes]).astype(np.float32))
export(net,
x,
label,
file_name="mindir/mini_alexnet_train",
file_format='MINDIR')
if len(sys.argv) > 1:
save_inout(sys.argv[1] + "mini_alexnet", x, label, n, net, sparse=False)
import sys
import numpy as np
from train_utils import save_inout, train_wrap
from official.cv.googlenet.src.googlenet import GoogleNet
import mindspore.common.dtype as mstype
from mindspore import context, Tensor, nn
from mindspore.train.serialization import export
context.set_context(mode=context.PYNATIVE_MODE,
device_target="GPU",
save_graphs=False)
n = GoogleNet(num_classes=10)
loss_fn = nn.SoftmaxCrossEntropyWithLogits(sparse=False)
optimizer = nn.SGD(n.trainable_params(),
learning_rate=0.01,
momentum=0.9,
dampening=0.0,
weight_decay=5e-4,
nesterov=True,
loss_scale=0.9)
net = train_wrap(n, loss_fn, optimizer)
batch = 2
x = Tensor(np.random.randn(batch, 3, 224, 224), mstype.float32)
label = Tensor(np.zeros([batch, 10]).astype(np.float32))
export(net, x, label, file_name="mindir/googlenet_train", file_format='MINDIR')
if len(sys.argv) > 1:
save_inout(sys.argv[1] + "googlenet", x, label, n, net)
import sys
import numpy as np
from train_utils import save_inout, train_wrap
from NetworkInNetwork import NiN
import mindspore.common.dtype as mstype
from mindspore import context, Tensor, nn
from mindspore.train.serialization import export
context.set_context(mode=context.PYNATIVE_MODE,
device_target="GPU",
save_graphs=False)
n = NiN(num_classes=10)
loss_fn = nn.SoftmaxCrossEntropyWithLogits(sparse=True, reduction="mean")
optimizer = nn.SGD(n.trainable_params(),
learning_rate=0.01,
momentum=0.9,
dampening=0.0,
weight_decay=5e-4,
nesterov=True,
loss_scale=0.9)
net = train_wrap(n, loss_fn, optimizer)
batch = 2
x = Tensor(np.random.randn(batch, 3, 32, 32), mstype.float32)
label = Tensor(np.zeros([batch]).astype(np.int32))
export(net, x, label, file_name="mindir/nin_train", file_format='MINDIR')
if len(sys.argv) > 1:
save_inout(sys.argv[1] + "nin", x, label, n, net)
from train_utils import save_inout, train_wrap
import mindspore.common.dtype as mstype
from mindspore import context, Tensor, nn
from mindspore.train.serialization import export
from src.network.densenet import DenseNet121
sys.path.append(os.environ['CLOUD_MODEL_ZOO'] + 'official/cv/densenet121/')
#pylint: disable=wrong-import-position
context.set_context(mode=context.PYNATIVE_MODE,
device_target="GPU",
save_graphs=False)
n = DenseNet121(num_classes=10)
loss_fn = nn.SoftmaxCrossEntropyWithLogits(sparse=False)
optimizer = nn.SGD(n.trainable_params(),
learning_rate=0.001,
momentum=0.9,
dampening=0.0,
weight_decay=0.0,
nesterov=True,
loss_scale=0.9)
net = train_wrap(n, loss_fn, optimizer)
batch = 2
x = Tensor(np.random.randn(batch, 3, 224, 224), mstype.float32)
label = Tensor(np.zeros([batch, 10]).astype(np.float32))
export(net, x, label, file_name="mindir/densenet_train", file_format='MINDIR')
if len(sys.argv) > 1:
save_inout(sys.argv[1] + "densenet", x, label, n, net)
# 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.
# ============================================================================
"""shufflenetv2_train_export."""
import sys
import numpy as np
from train_utils import save_inout, train_wrap
from official.cv.shufflenetv2.src.shufflenetv2 import ShuffleNetV2
import mindspore.common.dtype as mstype
from mindspore import context, Tensor, nn
from mindspore.train.serialization import export
context.set_context(mode=context.PYNATIVE_MODE, device_target="GPU", save_graphs=False)
n = ShuffleNetV2(n_class=10)
loss_fn = nn.SoftmaxCrossEntropyWithLogits(sparse=False)
optimizer = nn.Momentum(n.trainable_params(), 0.01, 0.9, use_nesterov=False)
net = train_wrap(n, loss_fn, optimizer)
batch = 2
x = Tensor(np.random.randn(batch, 3, 224, 224), mstype.float32)
label = Tensor(np.zeros([batch, 10]).astype(np.float32))
export(net, x, label, file_name="mindir/shufflenetv2_train", file_format='MINDIR')
if len(sys.argv) > 1:
save_inout(sys.argv[1] + "shufflenetv2", x, label, n, net)
from train_utils import save_inout, train_wrap
from official.cv.mobilenetv2.src.mobilenetV2 import MobileNetV2Backbone, MobileNetV2Head, mobilenet_v2
import mindspore.common.dtype as mstype
from mindspore import context, Tensor, nn
from mindspore.train.serialization import export
context.set_context(mode=context.PYNATIVE_MODE,
device_target="GPU",
save_graphs=False)
batch = 16
backbone_net = MobileNetV2Backbone()
head_net = MobileNetV2Head(input_channel=backbone_net.out_channels,
num_classes=10)
n = mobilenet_v2(backbone_net, head_net)
loss_fn = nn.SoftmaxCrossEntropyWithLogits(sparse=False)
optimizer = nn.Momentum(n.trainable_params(), 0.01, 0.9, use_nesterov=False)
net = train_wrap(n, loss_fn, optimizer)
x = Tensor(np.random.randn(batch, 3, 224, 224), mstype.float32)
label = Tensor(np.zeros([batch, 10]).astype(np.float32))
export(net,
x,
label,
file_name="mindir/mobilenetv2_train",
file_format='MINDIR')
if len(sys.argv) > 1:
save_inout(sys.argv[1] + "mobilenetv2", x, label, n, net, sparse=False)
from mindspore import context, Tensor, nn
from mindspore.train.serialization import export
context.set_context(mode=context.PYNATIVE_MODE,
device_target="GPU",
save_graphs=False)
n = MobileNetV1(10)
loss_fn = nn.SoftmaxCrossEntropyWithLogits(sparse=False)
optimizer = nn.SGD(n.trainable_params(),
learning_rate=1e-2,
momentum=0.9,
dampening=0.1,
weight_decay=0.0,
nesterov=False,
loss_scale=1.0)
net = train_wrap(n, loss_fn, optimizer)
batch = 2
x = Tensor(np.random.randn(batch, 3, 224, 224), mstype.float32)
label = Tensor(np.zeros([batch, 10]).astype(np.float32))
export(net,
x,
label,
file_name="mindir/mobilenetv1_train",
file_format='MINDIR')
if len(sys.argv) > 1:
save_inout(sys.argv[1] + "mobilenetv1", x, label, n, net)