Huawei_Technology
/
mindspore
Not watched
1
0
Fork 0
Code
Releases 13 Wiki evaluate Activity
Issues 0 Pull Requests 0 Datasets Model Cloudbrain HPC
Browse Source

delete platform in mobilenetV2.py

tags/v1.0.0
zhaoting 5 years ago
parent
6240189190
commit
4f801de34b
7 changed files with 29 additions and 65 deletions
Split View
Diff Options
Show Stats Download Patch File Download Diff File
  1. +8
    -24
      model_zoo/official/cv/mobilenetv2/Readme.md
  2. +3
    -7
      model_zoo/official/cv/mobilenetv2/eval.py
  3. +10
    -10
      model_zoo/official/cv/mobilenetv2/src/mobilenetV2.py
  4. +1
    -1
      model_zoo/official/cv/mobilenetv2/src/models.py
  5. +5
    -21
      model_zoo/official/cv/mobilenetv3/Readme.md
  6. +1
    -1
      model_zoo/official/cv/mobilenetv3/scripts/run_infer.sh
  7. +1
    -1
      model_zoo/official/cv/resnet50_quant/scripts/run_infer.sh

+ 8
- 24
model_zoo/official/cv/mobilenetv2/Readme.md View File

@@ -181,7 +181,7 @@ result: {'acc': 0.71976314102564111} ckpt=/path/to/checkpoint/mobilenet-200_625.

| Parameters | MobilenetV2 | |
| -------------------------- | ---------------------------------------------------------- | ------------------------- |
| Model Version | | large |
| Model Version | V1 | V1 |
| Resource | Ascend 910, cpu:2.60GHz 56cores, memory:314G | NV SMX2 V100-32G |
| uploaded Date | 05/06/2020 | 05/06/2020 |
| MindSpore Version | 0.3.0 | 0.3.0 |
@@ -189,29 +189,13 @@ result: {'acc': 0.71976314102564111} ckpt=/path/to/checkpoint/mobilenet-200_625.
| Training Parameters | src/config.py | src/config.py |
| Optimizer | Momentum | Momentum |
| Loss Function | SoftmaxCrossEntropy | SoftmaxCrossEntropy |
| outputs | | |
| Loss | | 1.913 |
| Accuracy | | ACC1[77.09%] ACC5[92.57%] |
| Total time | | |
| Params (M) | | |
| Checkpoint for Fine tuning | | |
| Model for inference | | |

#### Inference Performance

| Parameters | | | |
| -------------------------- | ----------------------------- | ------------------------- | -------------------- |
| Model Version | V1 | | |
| Resource | Ascend 910 | NV SMX2 V100-32G | Ascend 310 |
| uploaded Date | 05/06/2020 | 05/22/2020 | |
| MindSpore Version | 0.2.0 | 0.2.0 | 0.2.0 |
| Dataset | ImageNet, 1.2W | ImageNet, 1.2W | ImageNet, 1.2W |
| batch_size | | 130(8P) | |
| outputs | | | |
| Accuracy | | ACC1[72.07%] ACC5[90.90%] | |
| Speed | | | |
| Total time | | | |
| Model for inference | | | |
| outputs | probability | probability |
| Loss | 1.908 | 1.913 |
| Accuracy | ACC1[71.78%] | ACC1[71.08%] |
| Total time | 753 min | 845 min |
| Params (M) | 3.3 M | 3.3 M |
| Checkpoint for Fine tuning | 27.3 M | 27.3 M |
| Scripts | [Link](https://gitee.com/mindspore/mindspore/tree/master/model_zoo/official/cv/mobilenetv2)|

# [Description of Random Situation](#contents)



+ 3
- 7
model_zoo/official/cv/mobilenetv2/eval.py View File

@@ -21,18 +21,14 @@ from mindspore.common import dtype as mstype

from src.dataset import create_dataset
from src.config import set_config
from src.mobilenetV2 import MobileNetV2Backbone, MobileNetV2Head, mobilenet_v2
from src.args import eval_parse_args
from src.models import load_ckpt
from src.models import define_net, load_ckpt
from src.utils import switch_precision, set_context

if __name__ == '__main__':
args_opt = eval_parse_args()
config = set_config(args_opt)

backbone_net = MobileNetV2Backbone(platform=args_opt.platform)
head_net = MobileNetV2Head(input_channel=backbone_net.out_channels, num_classes=config.num_classes)
net = mobilenet_v2(backbone_net, head_net)
backbone_net, head_net, net = define_net(args_opt, config)

#load the trained checkpoint file to the net for evaluation
if args_opt.head_ckpt:
@@ -51,7 +47,7 @@ if __name__ == '__main__':
loss = nn.SoftmaxCrossEntropyWithLogits(sparse=True, reduction='mean')
model = Model(net, loss_fn=loss, metrics={'acc'})

res = model.eval(dataset, dataset_sink_mode=False)
res = model.eval(dataset)
print(f"result:{res}\npretrain_ckpt={args_opt.pretrain_ckpt}")
if args_opt.head_ckpt:
print(f"head_ckpt={args_opt.head_ckpt}")

+ 10
- 10
model_zoo/official/cv/mobilenetv2/src/mobilenetV2.py View File

@@ -72,7 +72,7 @@ class ConvBNReLU(nn.Cell):
>>> ConvBNReLU(16, 256, kernel_size=1, stride=1, groups=1)
"""

def __init__(self, platform, in_planes, out_planes, kernel_size=3, stride=1, groups=1):
def __init__(self, in_planes, out_planes, kernel_size=3, stride=1, groups=1):
super(ConvBNReLU, self).__init__()
padding = (kernel_size - 1) // 2
in_channels = in_planes
@@ -109,7 +109,7 @@ class InvertedResidual(nn.Cell):
>>> ResidualBlock(3, 256, 1, 1)
"""

def __init__(self, platform, inp, oup, stride, expand_ratio):
def __init__(self, inp, oup, stride, expand_ratio):
super(InvertedResidual, self).__init__()
assert stride in [1, 2]

@@ -118,10 +118,10 @@ class InvertedResidual(nn.Cell):

layers = []
if expand_ratio != 1:
layers.append(ConvBNReLU(platform, inp, hidden_dim, kernel_size=1))
layers.append(ConvBNReLU(inp, hidden_dim, kernel_size=1))
layers.extend([
# dw
ConvBNReLU(platform, hidden_dim, hidden_dim,
ConvBNReLU(hidden_dim, hidden_dim,
stride=stride, groups=hidden_dim),
# pw-linear
nn.Conv2d(hidden_dim, oup, kernel_size=1,
@@ -157,7 +157,7 @@ class MobileNetV2Backbone(nn.Cell):
>>> MobileNetV2(num_classes=1000)
"""

def __init__(self, platform, width_mult=1., inverted_residual_setting=None, round_nearest=8,
def __init__(self, width_mult=1., inverted_residual_setting=None, round_nearest=8,
input_channel=32, last_channel=1280):
super(MobileNetV2Backbone, self).__init__()
block = InvertedResidual
@@ -178,16 +178,16 @@ class MobileNetV2Backbone(nn.Cell):
# building first layer
input_channel = _make_divisible(input_channel * width_mult, round_nearest)
self.out_channels = _make_divisible(last_channel * max(1.0, width_mult), round_nearest)
features = [ConvBNReLU(platform, 3, input_channel, stride=2)]
features = [ConvBNReLU(3, input_channel, stride=2)]
# building inverted residual blocks
for t, c, n, s in self.cfgs:
output_channel = _make_divisible(c * width_mult, round_nearest)
for i in range(n):
stride = s if i == 0 else 1
features.append(block(platform, input_channel, output_channel, stride, expand_ratio=t))
features.append(block(input_channel, output_channel, stride, expand_ratio=t))
input_channel = output_channel
# building last several layers
features.append(ConvBNReLU(platform, input_channel, self.out_channels, kernel_size=1))
features.append(ConvBNReLU(input_channel, self.out_channels, kernel_size=1))
# make it nn.CellList
self.features = nn.SequentialCell(features)
self._initialize_weights()
@@ -293,10 +293,10 @@ class MobileNetV2(nn.Cell):
>>> MobileNetV2(backbone, head)
"""

def __init__(self, platform, num_classes=1000, width_mult=1., has_dropout=False, inverted_residual_setting=None, \
def __init__(self, num_classes=1000, width_mult=1., has_dropout=False, inverted_residual_setting=None, \
round_nearest=8, input_channel=32, last_channel=1280):
super(MobileNetV2, self).__init__()
self.backbone = MobileNetV2Backbone(platform=platform, width_mult=width_mult, \
self.backbone = MobileNetV2Backbone(width_mult=width_mult, \
inverted_residual_setting=inverted_residual_setting, \
round_nearest=round_nearest, input_channel=input_channel, last_channel=last_channel).get_features
self.head = MobileNetV2Head(input_channel=self.backbone.out_channel, num_classes=num_classes, \


+ 1
- 1
model_zoo/official/cv/mobilenetv2/src/models.py View File

@@ -120,7 +120,7 @@ def load_ckpt(network, pretrain_ckpt_path, trainable=True):
param.requires_grad = False

def define_net(args, config):
backbone_net = MobileNetV2Backbone(platform=args.platform)
backbone_net = MobileNetV2Backbone()
head_net = MobileNetV2Head(input_channel=backbone_net.out_channels, num_classes=config.num_classes)
net = mobilenet_v2(backbone_net, head_net)



+ 5
- 21
model_zoo/official/cv/mobilenetv3/Readme.md View File

@@ -145,29 +145,13 @@ result: {'acc': 0.71976314102564111} ckpt=/path/to/checkpoint/mobilenet-200_625.
| Training Parameters | src/config.py |
| Optimizer | Momentum |
| Loss Function | SoftmaxCrossEntropy |
| outputs | |
| outputs | probability |
| Loss | 1.913 |
| Accuracy | ACC1[77.57%] ACC5[92.51%] |
| Total time | |
| Params (M) | |
| Checkpoint for Fine tuning | |
| Model for inference | |

#### Inference Performance

| Parameters | |
| -------------------------- | -------------------- |
| Model Version | |
| Resource | NV SMX2 V100-32G |
| uploaded Date | 05/22/2020 |
| MindSpore Version | 0.2.0 |
| Dataset | ImageNet, 1.2W |
| batch_size | 130(8P) |
| outputs | |
| Accuracy | ACC1[75.43%] ACC5[92.51%] |
| Speed | |
| Total time | |
| Model for inference | |
| Total time | 1433 min |
| Params (M) | 5.48 M |
| Checkpoint for Fine tuning | 44 M |
| Scripts | [Link](https://gitee.com/mindspore/mindspore/tree/master/model_zoo/official/cv/mobilenetv3)|

# [Description of Random Situation](#contents)



+ 1
- 1
model_zoo/official/cv/mobilenetv3/scripts/run_infer.sh View File

@@ -46,7 +46,7 @@ fi
mkdir ../eval
cd ../eval || exit

# luanch
# launch
python ${BASEPATH}/../eval.py \
--device_target=$1 \
--dataset_path=$2 \


+ 1
- 1
model_zoo/official/cv/resnet50_quant/scripts/run_infer.sh View File

@@ -46,7 +46,7 @@ fi
mkdir ../eval
cd ../eval || exit

# luanch
# launch
python ${BASEPATH}/../eval.py \
--device_target=$1 \
--dataset_path=$2 \


Write Preview
Loading…
Cancel
Save