From d2846c0025af15750d1bc298d88fb0b861a01e96 Mon Sep 17 00:00:00 2001 From: Bingya Weng Date: Wed, 8 Sep 2021 10:51:02 +0800 Subject: [PATCH] modify test example of probability --- .../nn/probability/zhusuan/variational/elbo.py | 5 ++++- tests/st/ops/gpu/test_determinant_triangle.py | 2 +- tests/st/probability/bnn_layers/dataset.py | 2 +- tests/st/probability/bnn_layers/test_bnn_layer.py | 8 ++++---- tests/st/probability/dpn/test_gpu_svi_cvae.py | 8 ++++---- tests/st/probability/dpn/test_gpu_svi_vae.py | 6 +++--- tests/st/probability/dpn/test_gpu_vae_gan.py | 13 ++++++------- tests/st/probability/toolbox/test_uncertainty.py | 4 ++-- tests/st/probability/transforms/dataset.py | 2 +- .../transforms/test_transform_bnn_layer.py | 8 ++++---- .../transforms/test_transform_bnn_model.py | 8 ++++---- 11 files changed, 34 insertions(+), 32 deletions(-) diff --git a/mindspore/nn/probability/zhusuan/variational/elbo.py b/mindspore/nn/probability/zhusuan/variational/elbo.py index fd3bdd13eb1..8ba4a4eb410 100644 --- a/mindspore/nn/probability/zhusuan/variational/elbo.py +++ b/mindspore/nn/probability/zhusuan/variational/elbo.py @@ -32,9 +32,12 @@ class ELBO(nn.Cell): def construct(self, *inputs, **kwargs): if len(inputs) >= 2: x, y = inputs[0], inputs[1] - else: + elif len(inputs) >= 1: x = inputs[0] y = None + else: + x = None + y = None z, log_prob_z = self.variational(x, None, y) _, log_prob_x_, _, log_prob_z_ = self.generator(x, z, y) diff --git a/tests/st/ops/gpu/test_determinant_triangle.py b/tests/st/ops/gpu/test_determinant_triangle.py index 5460b68497b..961554832c7 100644 --- a/tests/st/ops/gpu/test_determinant_triangle.py +++ b/tests/st/ops/gpu/test_determinant_triangle.py @@ -32,7 +32,7 @@ class Net(nn.Cell): def construct(self, x): return self.det_triangle(x) -@pytest.mark.level1 +@pytest.mark.level2 @pytest.mark.platform_x86_gpu_training @pytest.mark.env_onecard def test_net_1D(): diff --git a/tests/st/probability/bnn_layers/dataset.py b/tests/st/probability/bnn_layers/dataset.py index df9eecda1fe..707410257c9 100644 --- a/tests/st/probability/bnn_layers/dataset.py +++ b/tests/st/probability/bnn_layers/dataset.py @@ -20,7 +20,7 @@ import mindspore.dataset as ds import mindspore.dataset.vision.c_transforms as CV import mindspore.dataset.transforms.c_transforms as C from mindspore.dataset.vision import Inter -from mindspore.common import dtype as mstype +from mindspore import dtype as mstype def create_dataset(data_path, batch_size=32, repeat_size=1, diff --git a/tests/st/probability/bnn_layers/test_bnn_layer.py b/tests/st/probability/bnn_layers/test_bnn_layer.py index a594ea83ed4..5be5100f881 100644 --- a/tests/st/probability/bnn_layers/test_bnn_layer.py +++ b/tests/st/probability/bnn_layers/test_bnn_layer.py @@ -20,7 +20,7 @@ from mindspore.common.initializer import TruncatedNormal import mindspore.nn as nn from mindspore.nn import TrainOneStepCell from mindspore.nn.probability import bnn_layers -from mindspore.ops import operations as P +import mindspore.ops as ops from mindspore import context from dataset import create_dataset @@ -71,7 +71,7 @@ class BNNLeNet5(nn.Cell): self.relu = nn.ReLU() self.max_pool2d = nn.MaxPool2d(kernel_size=2, stride=2) self.flatten = nn.Flatten() - self.reshape = P.Reshape() + self.reshape = ops.Reshape() def construct(self, x): x = self.conv1(x) @@ -97,7 +97,7 @@ def train_model(train_net, net, dataset): label = Tensor(data['label'].astype(np.int32)) loss = train_net(train_x, label) output = net(train_x) - log_output = P.LogSoftmax(axis=1)(output) + log_output = ops.LogSoftmax(axis=1)(output) acc = np.mean(log_output.asnumpy().argmax(axis=1) == label.asnumpy()) accs.append(acc) loss_sum += loss.asnumpy() @@ -113,7 +113,7 @@ def validate_model(net, dataset): train_x = Tensor(data['image'].astype(np.float32)) label = Tensor(data['label'].astype(np.int32)) output = net(train_x) - log_output = P.LogSoftmax(axis=1)(output) + log_output = ops.LogSoftmax(axis=1)(output) acc = np.mean(log_output.asnumpy().argmax(axis=1) == label.asnumpy()) accs.append(acc) diff --git a/tests/st/probability/dpn/test_gpu_svi_cvae.py b/tests/st/probability/dpn/test_gpu_svi_cvae.py index fbd8d346afb..1c08a26af96 100644 --- a/tests/st/probability/dpn/test_gpu_svi_cvae.py +++ b/tests/st/probability/dpn/test_gpu_svi_cvae.py @@ -14,12 +14,12 @@ # ============================================================================ import os -import mindspore.common.dtype as mstype +from mindspore import dtype as mstype import mindspore.dataset as ds import mindspore.dataset.vision.c_transforms as CV import mindspore.nn as nn from mindspore import context, Tensor -from mindspore.ops import operations as P +import mindspore.ops as ops from mindspore.nn.probability.dpn import ConditionalVAE from mindspore.nn.probability.infer import ELBO, SVI @@ -34,7 +34,7 @@ class Encoder(nn.Cell): self.fc1 = nn.Dense(1024 + num_classes, 400) self.relu = nn.ReLU() self.flatten = nn.Flatten() - self.concat = P.Concat(axis=1) + self.concat = ops.Concat(axis=1) self.one_hot = nn.OneHot(depth=num_classes) def construct(self, x, y): @@ -51,7 +51,7 @@ class Decoder(nn.Cell): super(Decoder, self).__init__() self.fc2 = nn.Dense(400, 1024) self.sigmoid = nn.Sigmoid() - self.reshape = P.Reshape() + self.reshape = ops.Reshape() def construct(self, z): z = self.fc2(z) diff --git a/tests/st/probability/dpn/test_gpu_svi_vae.py b/tests/st/probability/dpn/test_gpu_svi_vae.py index e0a9001139f..6a84bed90e6 100644 --- a/tests/st/probability/dpn/test_gpu_svi_vae.py +++ b/tests/st/probability/dpn/test_gpu_svi_vae.py @@ -14,12 +14,12 @@ # ============================================================================ import os -import mindspore.common.dtype as mstype +from mindspore import dtype as mstype import mindspore.dataset as ds import mindspore.dataset.vision.c_transforms as CV import mindspore.nn as nn from mindspore import context, Tensor -from mindspore.ops import operations as P +import mindspore.ops as ops from mindspore.nn.probability.dpn import VAE from mindspore.nn.probability.infer import ELBO, SVI @@ -50,7 +50,7 @@ class Decoder(nn.Cell): super(Decoder, self).__init__() self.fc1 = nn.Dense(400, 1024) self.sigmoid = nn.Sigmoid() - self.reshape = P.Reshape() + self.reshape = ops.Reshape() def construct(self, z): z = self.fc1(z) diff --git a/tests/st/probability/dpn/test_gpu_vae_gan.py b/tests/st/probability/dpn/test_gpu_vae_gan.py index ac0eeb07ff9..a606bd12ad7 100644 --- a/tests/st/probability/dpn/test_gpu_vae_gan.py +++ b/tests/st/probability/dpn/test_gpu_vae_gan.py @@ -21,8 +21,7 @@ import mindspore.dataset as ds import mindspore.dataset.vision.c_transforms as CV import mindspore.nn as nn from mindspore import context -from mindspore.ops import operations as P -from mindspore.ops import composite as C +import mindspore.ops as ops from mindspore.nn.probability.dpn import VAE from mindspore.nn.probability.infer import ELBO, SVI @@ -51,7 +50,7 @@ class Decoder(nn.Cell): self.fc1 = nn.Dense(400, 1024) self.relu = nn.ReLU() self.sigmoid = nn.Sigmoid() - self.reshape = P.Reshape() + self.reshape = ops.Reshape() def construct(self, z): z = self.fc1(z) @@ -93,9 +92,9 @@ class VaeGan(nn.Cell): self.D = Discriminator() self.dense = nn.Dense(20, 400) self.vae = VAE(self.E, self.G, 400, 20) - self.shape = P.Shape() - self.normal = C.normal - self.to_tensor = P.ScalarToArray() + self.shape = ops.Shape() + self.normal = ops.normal + self.to_tensor = ops.ScalarToArray() def construct(self, x): recon_x, x, mu, std = self.vae(x) @@ -111,7 +110,7 @@ class VaeGan(nn.Cell): class VaeGanLoss(ELBO): def __init__(self): super(VaeGanLoss, self).__init__() - self.zeros = P.ZerosLike() + self.zeros = ops.ZerosLike() self.mse = nn.MSELoss(reduction='sum') def construct(self, data, label): diff --git a/tests/st/probability/toolbox/test_uncertainty.py b/tests/st/probability/toolbox/test_uncertainty.py index 4b86e2fab0d..b406ad31cc6 100644 --- a/tests/st/probability/toolbox/test_uncertainty.py +++ b/tests/st/probability/toolbox/test_uncertainty.py @@ -18,11 +18,11 @@ import mindspore.dataset.transforms.c_transforms as C import mindspore.dataset.vision.c_transforms as CV import mindspore.nn as nn from mindspore import context, Tensor -from mindspore.common import dtype as mstype +from mindspore import dtype as mstype from mindspore.common.initializer import TruncatedNormal from mindspore.dataset.vision import Inter from mindspore.nn.probability.toolbox.uncertainty_evaluation import UncertaintyEvaluation -from mindspore.train.serialization import load_checkpoint, load_param_into_net +from mindspore.train import load_checkpoint, load_param_into_net context.set_context(mode=context.GRAPH_MODE, device_target="GPU") diff --git a/tests/st/probability/transforms/dataset.py b/tests/st/probability/transforms/dataset.py index df9eecda1fe..707410257c9 100644 --- a/tests/st/probability/transforms/dataset.py +++ b/tests/st/probability/transforms/dataset.py @@ -20,7 +20,7 @@ import mindspore.dataset as ds import mindspore.dataset.vision.c_transforms as CV import mindspore.dataset.transforms.c_transforms as C from mindspore.dataset.vision import Inter -from mindspore.common import dtype as mstype +from mindspore import dtype as mstype def create_dataset(data_path, batch_size=32, repeat_size=1, diff --git a/tests/st/probability/transforms/test_transform_bnn_layer.py b/tests/st/probability/transforms/test_transform_bnn_layer.py index cb28b19f92f..35c0e9d599e 100644 --- a/tests/st/probability/transforms/test_transform_bnn_layer.py +++ b/tests/st/probability/transforms/test_transform_bnn_layer.py @@ -20,7 +20,7 @@ from mindspore.common.initializer import TruncatedNormal import mindspore.nn as nn from mindspore.nn import TrainOneStepCell, WithLossCell from mindspore.nn.probability import transforms, bnn_layers -from mindspore.ops import operations as P +import mindspore.ops as ops from mindspore import context from dataset import create_dataset @@ -72,7 +72,7 @@ class LeNet5(nn.Cell): self.relu = nn.ReLU() self.max_pool2d = nn.MaxPool2d(kernel_size=2, stride=2) self.flatten = nn.Flatten() - self.reshape = P.Reshape() + self.reshape = ops.Reshape() def construct(self, x): x = self.conv1(x) @@ -98,7 +98,7 @@ def train_model(train_net, net, dataset): label = Tensor(data['label'].astype(np.int32)) loss = train_net(train_x, label) output = net(train_x) - log_output = P.LogSoftmax(axis=1)(output) + log_output = ops.LogSoftmax(axis=1)(output) acc = np.mean(log_output.asnumpy().argmax(axis=1) == label.asnumpy()) accs.append(acc) loss_sum += loss.asnumpy() @@ -114,7 +114,7 @@ def validate_model(net, dataset): train_x = Tensor(data['image'].astype(np.float32)) label = Tensor(data['label'].astype(np.int32)) output = net(train_x) - log_output = P.LogSoftmax(axis=1)(output) + log_output = ops.LogSoftmax(axis=1)(output) acc = np.mean(log_output.asnumpy().argmax(axis=1) == label.asnumpy()) accs.append(acc) diff --git a/tests/st/probability/transforms/test_transform_bnn_model.py b/tests/st/probability/transforms/test_transform_bnn_model.py index d83033e4a08..b03cf173e0f 100644 --- a/tests/st/probability/transforms/test_transform_bnn_model.py +++ b/tests/st/probability/transforms/test_transform_bnn_model.py @@ -19,7 +19,7 @@ from mindspore.common.initializer import TruncatedNormal import mindspore.nn as nn from mindspore.nn import WithLossCell, TrainOneStepCell from mindspore.nn.probability import transforms -from mindspore.ops import operations as P +import mindspore.ops as ops from mindspore import context from dataset import create_dataset @@ -71,7 +71,7 @@ class LeNet5(nn.Cell): self.relu = nn.ReLU() self.max_pool2d = nn.MaxPool2d(kernel_size=2, stride=2) self.flatten = nn.Flatten() - self.reshape = P.Reshape() + self.reshape = ops.Reshape() def construct(self, x): x = self.conv1(x) @@ -97,7 +97,7 @@ def train_model(train_net, net, dataset): label = Tensor(data['label'].astype(np.int32)) loss = train_net(train_x, label) output = net(train_x) - log_output = P.LogSoftmax(axis=1)(output) + log_output = ops.LogSoftmax(axis=1)(output) acc = np.mean(log_output.asnumpy().argmax(axis=1) == label.asnumpy()) accs.append(acc) loss_sum += loss.asnumpy() @@ -113,7 +113,7 @@ def validate_model(net, dataset): train_x = Tensor(data['image'].astype(np.float32)) label = Tensor(data['label'].astype(np.int32)) output = net(train_x) - log_output = P.LogSoftmax(axis=1)(output) + log_output = ops.LogSoftmax(axis=1)(output) acc = np.mean(log_output.asnumpy().argmax(axis=1) == label.asnumpy()) accs.append(acc)