PyTorch Lightning 1.1: research : CIFAR100 (MobileNet)
作成 : (株)クラスキャット セールスインフォメーション
作成日時 : 02/25/2021 (1.1.x)
* 本ページは以下の CIFAR10 用リソースを参考に CIFAR100 で遂行した実験結果のレポートです：
- notebooks : PyTorch Lightning CIFAR10 ~94% Baseline Tutorial
- Train CIFAR10 with PyTorch
- Pytorch-cifar100
* ご自由にリンクを張って頂いてかまいませんが、sales-info@classcat.com までご一報いただけると嬉しいです。
★ 無料セミナー実施中 ★ クラスキャット主催 人工知能 & ビジネス Web セミナー
人工知能とビジネスをテーマにウェビナー (WEB セミナー) を定期的に開催しています。スケジュールは弊社 公式 Web サイト でご確認頂けます。
- お住まいの地域に関係なく Web ブラウザからご参加頂けます。事前登録 が必要ですのでご注意ください。
- Windows PC のブラウザからご参加が可能です。スマートデバイスもご利用可能です。
クラスキャットは人工知能・テレワークに関する各種サービスを提供しております :
| 人工知能研究開発支援 | 人工知能研修サービス | テレワーク & オンライン授業を支援 |
| PoC(概念実証)を失敗させないための支援 (本支援はセミナーに参加しアンケートに回答した方を対象としています。) | ||
◆ お問合せ : 本件に関するお問い合わせ先は下記までお願いいたします。
| 株式会社クラスキャット セールス・マーケティング本部 セールス・インフォメーション |
| E-Mail：sales-info@classcat.com ; WebSite: https://www.classcat.com/ |
| Facebook: https://www.facebook.com/ClassCatJP/ |
research: CIFAR100 (MobileNet)
仕様
- Total params: 2,369,380 (2.4M)
- Trainable params: 2,369,380
- Non-trainable params: 0
結果
- MobileNetV2
- {‘test_acc’: 0.6323999762535095, ‘test_loss’: 1.4939695596694946}
- 100 エポック ; Wall time: 1h 20min 26s
- Tesla T4
- ReduceLROnPlateau
CIFAR 100 DM
from typing import Any, Callable, Optional, Sequence, Union
from pl_bolts.datamodules.vision_datamodule import VisionDataModule
#from pl_bolts.datasets import TrialCIFAR10
#from pl_bolts.transforms.dataset_normalizations import cifar10_normalization
from pl_bolts.utils import _TORCHVISION_AVAILABLE
from pl_bolts.utils.warnings import warn_missing_pkg
if _TORCHVISION_AVAILABLE:
from torchvision import transforms
#from torchvision import transforms as transform_lib
from torchvision.datasets import CIFAR100
else: # pragma: no cover
warn_missing_pkg('torchvision')
CIFAR100 = None
def cifar100_normalization():
if not _TORCHVISION_AVAILABLE: # pragma: no cover
raise ModuleNotFoundError(
'You want to use `torchvision` which is not installed yet, install it with `pip install torchvision`.'
)
normalize = transforms.Normalize(
mean=[x / 255.0 for x in [129.3, 124.1, 112.4]],
std=[x / 255.0 for x in [68.2, 65.4, 70.4]],
# cifar10
#mean=[x / 255.0 for x in [125.3, 123.0, 113.9]],
#std=[x / 255.0 for x in [63.0, 62.1, 66.7]],
)
return normalize
class CIFAR100DataModule(VisionDataModule):
"""
.. figure:: https://3qeqpr26caki16dnhd19sv6by6v-wpengine.netdna-ssl.com/wp-content/uploads/2019/01/
Plot-of-a-Subset-of-Images-from-the-CIFAR-10-Dataset.png
:width: 400
:alt: CIFAR-10
Specs:
- 10 classes (1 per class)
- Each image is (3 x 32 x 32)
Standard CIFAR10, train, val, test splits and transforms
Transforms::
mnist_transforms = transform_lib.Compose([
transform_lib.ToTensor(),
transforms.Normalize(
mean=[x / 255.0 for x in [125.3, 123.0, 113.9]],
std=[x / 255.0 for x in [63.0, 62.1, 66.7]]
)
])
Example::
from pl_bolts.datamodules import CIFAR10DataModule
dm = CIFAR10DataModule(PATH)
model = LitModel()
Trainer().fit(model, datamodule=dm)
Or you can set your own transforms
Example::
dm.train_transforms = ...
dm.test_transforms = ...
dm.val_transforms = ...
"""
name = "cifar100"
dataset_cls = CIFAR100
dims = (3, 32, 32)
def __init__(
self,
data_dir: Optional[str] = None,
val_split: Union[int, float] = 0.2,
num_workers: int = 16,
normalize: bool = False,
batch_size: int = 32,
seed: int = 42,
shuffle: bool = False,
pin_memory: bool = False,
drop_last: bool = False,
*args: Any,
**kwargs: Any,
) -> None:
"""
Args:
data_dir: Where to save/load the data
val_split: Percent (float) or number (int) of samples to use for the validation split
num_workers: How many workers to use for loading data
normalize: If true applies image normalize
batch_size: How many samples per batch to load
seed: Random seed to be used for train/val/test splits
shuffle: If true shuffles the train data every epoch
pin_memory: If true, the data loader will copy Tensors into CUDA pinned memory before
returning them
drop_last: If true drops the last incomplete batch
"""
super().__init__( # type: ignore[misc]
data_dir=data_dir,
val_split=val_split,
num_workers=num_workers,
normalize=normalize,
batch_size=batch_size,
seed=seed,
shuffle=shuffle,
pin_memory=pin_memory,
drop_last=drop_last,
*args,
**kwargs,
)
@property
def num_samples(self) -> int:
train_len, _ = self._get_splits(len_dataset=50_000)
return train_len
@property
def num_classes(self) -> int:
"""
Return:
10
"""
return 100
def default_transforms(self) -> Callable:
if self.normalize:
cf100_transforms = transforms.Compose([transform_lib.ToTensor(), cifar100_normalization()])
else:
cf100_transforms = transforms.Compose([transform_lib.ToTensor()])
return cf100_transforms
モデル
import torch import torch.nn as nn import torch.nn.functional as F
class LinearBottleNeck(nn.Module):
def __init__(self, in_channels, out_channels, stride, t=6, class_num=100):
super().__init__()
self.residual = nn.Sequential(
nn.Conv2d(in_channels, in_channels * t, 1),
nn.BatchNorm2d(in_channels * t),
nn.ReLU6(inplace=True),
nn.Conv2d(in_channels * t, in_channels * t, 3, stride=stride, padding=1, groups=in_channels * t),
nn.BatchNorm2d(in_channels * t),
nn.ReLU6(inplace=True),
nn.Conv2d(in_channels * t, out_channels, 1),
nn.BatchNorm2d(out_channels)
)
self.stride = stride
self.in_channels = in_channels
self.out_channels = out_channels
def forward(self, x):
residual = self.residual(x)
if self.stride == 1 and self.in_channels == self.out_channels:
residual += x
return residual
class MobileNetV2(nn.Module):
def __init__(self, class_num=100):
super().__init__()
self.pre = nn.Sequential(
nn.Conv2d(3, 32, 1, padding=1),
nn.BatchNorm2d(32),
nn.ReLU6(inplace=True)
)
self.stage1 = LinearBottleNeck(32, 16, 1, 1)
self.stage2 = self._make_stage(2, 16, 24, 2, 6)
self.stage3 = self._make_stage(3, 24, 32, 2, 6)
self.stage4 = self._make_stage(4, 32, 64, 2, 6)
self.stage5 = self._make_stage(3, 64, 96, 1, 6)
self.stage6 = self._make_stage(3, 96, 160, 1, 6)
self.stage7 = LinearBottleNeck(160, 320, 1, 6)
self.conv1 = nn.Sequential(
nn.Conv2d(320, 1280, 1),
nn.BatchNorm2d(1280),
nn.ReLU6(inplace=True)
)
self.conv2 = nn.Conv2d(1280, class_num, 1)
def forward(self, x):
x = self.pre(x)
x = self.stage1(x)
x = self.stage2(x)
x = self.stage3(x)
x = self.stage4(x)
x = self.stage5(x)
x = self.stage6(x)
x = self.stage7(x)
x = self.conv1(x)
x = F.adaptive_avg_pool2d(x, 1)
x = self.conv2(x)
x = x.view(x.size(0), -1)
return x
def _make_stage(self, repeat, in_channels, out_channels, stride, t):
layers = []
layers.append(LinearBottleNeck(in_channels, out_channels, stride, t))
while repeat - 1:
layers.append(LinearBottleNeck(out_channels, out_channels, 1, t))
repeat -= 1
return nn.Sequential(*layers)
def mobilenetv2():
return MobileNetV2()
net = mobilenetv2() print(net) y = net(torch.randn(1, 3, 32, 32)) print(y.size())
MobileNetV2(
(pre): Sequential(
(0): Conv2d(3, 32, kernel_size=(1, 1), stride=(1, 1), padding=(1, 1))
(1): BatchNorm2d(32, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(2): ReLU6(inplace=True)
)
(stage1): LinearBottleNeck(
(residual): Sequential(
(0): Conv2d(32, 32, kernel_size=(1, 1), stride=(1, 1))
(1): BatchNorm2d(32, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(2): ReLU6(inplace=True)
(3): Conv2d(32, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), groups=32)
(4): BatchNorm2d(32, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(5): ReLU6(inplace=True)
(6): Conv2d(32, 16, kernel_size=(1, 1), stride=(1, 1))
(7): BatchNorm2d(16, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
)
)
(stage2): Sequential(
(0): LinearBottleNeck(
(residual): Sequential(
(0): Conv2d(16, 96, kernel_size=(1, 1), stride=(1, 1))
(1): BatchNorm2d(96, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(2): ReLU6(inplace=True)
(3): Conv2d(96, 96, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), groups=96)
(4): BatchNorm2d(96, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(5): ReLU6(inplace=True)
(6): Conv2d(96, 24, kernel_size=(1, 1), stride=(1, 1))
(7): BatchNorm2d(24, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
)
)
(1): LinearBottleNeck(
(residual): Sequential(
(0): Conv2d(24, 144, kernel_size=(1, 1), stride=(1, 1))
(1): BatchNorm2d(144, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(2): ReLU6(inplace=True)
(3): Conv2d(144, 144, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), groups=144)
(4): BatchNorm2d(144, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(5): ReLU6(inplace=True)
(6): Conv2d(144, 24, kernel_size=(1, 1), stride=(1, 1))
(7): BatchNorm2d(24, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
)
)
)
(stage3): Sequential(
(0): LinearBottleNeck(
(residual): Sequential(
(0): Conv2d(24, 144, kernel_size=(1, 1), stride=(1, 1))
(1): BatchNorm2d(144, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(2): ReLU6(inplace=True)
(3): Conv2d(144, 144, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), groups=144)
(4): BatchNorm2d(144, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(5): ReLU6(inplace=True)
(6): Conv2d(144, 32, kernel_size=(1, 1), stride=(1, 1))
(7): BatchNorm2d(32, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
)
)
(1): LinearBottleNeck(
(residual): Sequential(
(0): Conv2d(32, 192, kernel_size=(1, 1), stride=(1, 1))
(1): BatchNorm2d(192, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(2): ReLU6(inplace=True)
(3): Conv2d(192, 192, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), groups=192)
(4): BatchNorm2d(192, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(5): ReLU6(inplace=True)
(6): Conv2d(192, 32, kernel_size=(1, 1), stride=(1, 1))
(7): BatchNorm2d(32, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
)
)
(2): LinearBottleNeck(
(residual): Sequential(
(0): Conv2d(32, 192, kernel_size=(1, 1), stride=(1, 1))
(1): BatchNorm2d(192, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(2): ReLU6(inplace=True)
(3): Conv2d(192, 192, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), groups=192)
(4): BatchNorm2d(192, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(5): ReLU6(inplace=True)
(6): Conv2d(192, 32, kernel_size=(1, 1), stride=(1, 1))
(7): BatchNorm2d(32, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
)
)
)
(stage4): Sequential(
(0): LinearBottleNeck(
(residual): Sequential(
(0): Conv2d(32, 192, kernel_size=(1, 1), stride=(1, 1))
(1): BatchNorm2d(192, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(2): ReLU6(inplace=True)
(3): Conv2d(192, 192, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), groups=192)
(4): BatchNorm2d(192, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(5): ReLU6(inplace=True)
(6): Conv2d(192, 64, kernel_size=(1, 1), stride=(1, 1))
(7): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
)
)
(1): LinearBottleNeck(
(residual): Sequential(
(0): Conv2d(64, 384, kernel_size=(1, 1), stride=(1, 1))
(1): BatchNorm2d(384, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(2): ReLU6(inplace=True)
(3): Conv2d(384, 384, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), groups=384)
(4): BatchNorm2d(384, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(5): ReLU6(inplace=True)
(6): Conv2d(384, 64, kernel_size=(1, 1), stride=(1, 1))
(7): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
)
)
(2): LinearBottleNeck(
(residual): Sequential(
(0): Conv2d(64, 384, kernel_size=(1, 1), stride=(1, 1))
(1): BatchNorm2d(384, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(2): ReLU6(inplace=True)
(3): Conv2d(384, 384, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), groups=384)
(4): BatchNorm2d(384, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(5): ReLU6(inplace=True)
(6): Conv2d(384, 64, kernel_size=(1, 1), stride=(1, 1))
(7): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
)
)
(3): LinearBottleNeck(
(residual): Sequential(
(0): Conv2d(64, 384, kernel_size=(1, 1), stride=(1, 1))
(1): BatchNorm2d(384, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(2): ReLU6(inplace=True)
(3): Conv2d(384, 384, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), groups=384)
(4): BatchNorm2d(384, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(5): ReLU6(inplace=True)
(6): Conv2d(384, 64, kernel_size=(1, 1), stride=(1, 1))
(7): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
)
)
)
(stage5): Sequential(
(0): LinearBottleNeck(
(residual): Sequential(
(0): Conv2d(64, 384, kernel_size=(1, 1), stride=(1, 1))
(1): BatchNorm2d(384, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(2): ReLU6(inplace=True)
(3): Conv2d(384, 384, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), groups=384)
(4): BatchNorm2d(384, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(5): ReLU6(inplace=True)
(6): Conv2d(384, 96, kernel_size=(1, 1), stride=(1, 1))
(7): BatchNorm2d(96, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
)
)
(1): LinearBottleNeck(
(residual): Sequential(
(0): Conv2d(96, 576, kernel_size=(1, 1), stride=(1, 1))
(1): BatchNorm2d(576, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(2): ReLU6(inplace=True)
(3): Conv2d(576, 576, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), groups=576)
(4): BatchNorm2d(576, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(5): ReLU6(inplace=True)
(6): Conv2d(576, 96, kernel_size=(1, 1), stride=(1, 1))
(7): BatchNorm2d(96, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
)
)
(2): LinearBottleNeck(
(residual): Sequential(
(0): Conv2d(96, 576, kernel_size=(1, 1), stride=(1, 1))
(1): BatchNorm2d(576, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(2): ReLU6(inplace=True)
(3): Conv2d(576, 576, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), groups=576)
(4): BatchNorm2d(576, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(5): ReLU6(inplace=True)
(6): Conv2d(576, 96, kernel_size=(1, 1), stride=(1, 1))
(7): BatchNorm2d(96, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
)
)
)
(stage6): Sequential(
(0): LinearBottleNeck(
(residual): Sequential(
(0): Conv2d(96, 576, kernel_size=(1, 1), stride=(1, 1))
(1): BatchNorm2d(576, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(2): ReLU6(inplace=True)
(3): Conv2d(576, 576, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), groups=576)
(4): BatchNorm2d(576, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(5): ReLU6(inplace=True)
(6): Conv2d(576, 160, kernel_size=(1, 1), stride=(1, 1))
(7): BatchNorm2d(160, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
)
)
(1): LinearBottleNeck(
(residual): Sequential(
(0): Conv2d(160, 960, kernel_size=(1, 1), stride=(1, 1))
(1): BatchNorm2d(960, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(2): ReLU6(inplace=True)
(3): Conv2d(960, 960, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), groups=960)
(4): BatchNorm2d(960, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(5): ReLU6(inplace=True)
(6): Conv2d(960, 160, kernel_size=(1, 1), stride=(1, 1))
(7): BatchNorm2d(160, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
)
)
(2): LinearBottleNeck(
(residual): Sequential(
(0): Conv2d(160, 960, kernel_size=(1, 1), stride=(1, 1))
(1): BatchNorm2d(960, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(2): ReLU6(inplace=True)
(3): Conv2d(960, 960, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), groups=960)
(4): BatchNorm2d(960, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(5): ReLU6(inplace=True)
(6): Conv2d(960, 160, kernel_size=(1, 1), stride=(1, 1))
(7): BatchNorm2d(160, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
)
)
)
(stage7): LinearBottleNeck(
(residual): Sequential(
(0): Conv2d(160, 960, kernel_size=(1, 1), stride=(1, 1))
(1): BatchNorm2d(960, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(2): ReLU6(inplace=True)
(3): Conv2d(960, 960, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), groups=960)
(4): BatchNorm2d(960, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(5): ReLU6(inplace=True)
(6): Conv2d(960, 320, kernel_size=(1, 1), stride=(1, 1))
(7): BatchNorm2d(320, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
)
)
(conv1): Sequential(
(0): Conv2d(320, 1280, kernel_size=(1, 1), stride=(1, 1))
(1): BatchNorm2d(1280, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(2): ReLU6(inplace=True)
)
(conv2): Conv2d(1280, 100, kernel_size=(1, 1), stride=(1, 1))
)
torch.Size([1, 100])
from torchsummary import summary
summary(mobilenetv2().to('cuda'), (3, 32, 32))
----------------------------------------------------------------
Layer (type) Output Shape Param #
================================================================
Conv2d-1 [-1, 32, 34, 34] 128
BatchNorm2d-2 [-1, 32, 34, 34] 64
ReLU6-3 [-1, 32, 34, 34] 0
Conv2d-4 [-1, 32, 34, 34] 1,056
BatchNorm2d-5 [-1, 32, 34, 34] 64
ReLU6-6 [-1, 32, 34, 34] 0
Conv2d-7 [-1, 32, 34, 34] 320
BatchNorm2d-8 [-1, 32, 34, 34] 64
ReLU6-9 [-1, 32, 34, 34] 0
Conv2d-10 [-1, 16, 34, 34] 528
BatchNorm2d-11 [-1, 16, 34, 34] 32
LinearBottleNeck-12 [-1, 16, 34, 34] 0
Conv2d-13 [-1, 96, 34, 34] 1,632
BatchNorm2d-14 [-1, 96, 34, 34] 192
ReLU6-15 [-1, 96, 34, 34] 0
Conv2d-16 [-1, 96, 17, 17] 960
BatchNorm2d-17 [-1, 96, 17, 17] 192
ReLU6-18 [-1, 96, 17, 17] 0
Conv2d-19 [-1, 24, 17, 17] 2,328
BatchNorm2d-20 [-1, 24, 17, 17] 48
LinearBottleNeck-21 [-1, 24, 17, 17] 0
Conv2d-22 [-1, 144, 17, 17] 3,600
BatchNorm2d-23 [-1, 144, 17, 17] 288
ReLU6-24 [-1, 144, 17, 17] 0
Conv2d-25 [-1, 144, 17, 17] 1,440
BatchNorm2d-26 [-1, 144, 17, 17] 288
ReLU6-27 [-1, 144, 17, 17] 0
Conv2d-28 [-1, 24, 17, 17] 3,480
BatchNorm2d-29 [-1, 24, 17, 17] 48
LinearBottleNeck-30 [-1, 24, 17, 17] 0
Conv2d-31 [-1, 144, 17, 17] 3,600
BatchNorm2d-32 [-1, 144, 17, 17] 288
ReLU6-33 [-1, 144, 17, 17] 0
Conv2d-34 [-1, 144, 9, 9] 1,440
BatchNorm2d-35 [-1, 144, 9, 9] 288
ReLU6-36 [-1, 144, 9, 9] 0
Conv2d-37 [-1, 32, 9, 9] 4,640
BatchNorm2d-38 [-1, 32, 9, 9] 64
LinearBottleNeck-39 [-1, 32, 9, 9] 0
Conv2d-40 [-1, 192, 9, 9] 6,336
BatchNorm2d-41 [-1, 192, 9, 9] 384
ReLU6-42 [-1, 192, 9, 9] 0
Conv2d-43 [-1, 192, 9, 9] 1,920
BatchNorm2d-44 [-1, 192, 9, 9] 384
ReLU6-45 [-1, 192, 9, 9] 0
Conv2d-46 [-1, 32, 9, 9] 6,176
BatchNorm2d-47 [-1, 32, 9, 9] 64
LinearBottleNeck-48 [-1, 32, 9, 9] 0
Conv2d-49 [-1, 192, 9, 9] 6,336
BatchNorm2d-50 [-1, 192, 9, 9] 384
ReLU6-51 [-1, 192, 9, 9] 0
Conv2d-52 [-1, 192, 9, 9] 1,920
BatchNorm2d-53 [-1, 192, 9, 9] 384
ReLU6-54 [-1, 192, 9, 9] 0
Conv2d-55 [-1, 32, 9, 9] 6,176
BatchNorm2d-56 [-1, 32, 9, 9] 64
LinearBottleNeck-57 [-1, 32, 9, 9] 0
Conv2d-58 [-1, 192, 9, 9] 6,336
BatchNorm2d-59 [-1, 192, 9, 9] 384
ReLU6-60 [-1, 192, 9, 9] 0
Conv2d-61 [-1, 192, 5, 5] 1,920
BatchNorm2d-62 [-1, 192, 5, 5] 384
ReLU6-63 [-1, 192, 5, 5] 0
Conv2d-64 [-1, 64, 5, 5] 12,352
BatchNorm2d-65 [-1, 64, 5, 5] 128
LinearBottleNeck-66 [-1, 64, 5, 5] 0
Conv2d-67 [-1, 384, 5, 5] 24,960
BatchNorm2d-68 [-1, 384, 5, 5] 768
ReLU6-69 [-1, 384, 5, 5] 0
Conv2d-70 [-1, 384, 5, 5] 3,840
BatchNorm2d-71 [-1, 384, 5, 5] 768
ReLU6-72 [-1, 384, 5, 5] 0
Conv2d-73 [-1, 64, 5, 5] 24,640
BatchNorm2d-74 [-1, 64, 5, 5] 128
LinearBottleNeck-75 [-1, 64, 5, 5] 0
Conv2d-76 [-1, 384, 5, 5] 24,960
BatchNorm2d-77 [-1, 384, 5, 5] 768
ReLU6-78 [-1, 384, 5, 5] 0
Conv2d-79 [-1, 384, 5, 5] 3,840
BatchNorm2d-80 [-1, 384, 5, 5] 768
ReLU6-81 [-1, 384, 5, 5] 0
Conv2d-82 [-1, 64, 5, 5] 24,640
BatchNorm2d-83 [-1, 64, 5, 5] 128
LinearBottleNeck-84 [-1, 64, 5, 5] 0
Conv2d-85 [-1, 384, 5, 5] 24,960
BatchNorm2d-86 [-1, 384, 5, 5] 768
ReLU6-87 [-1, 384, 5, 5] 0
Conv2d-88 [-1, 384, 5, 5] 3,840
BatchNorm2d-89 [-1, 384, 5, 5] 768
ReLU6-90 [-1, 384, 5, 5] 0
Conv2d-91 [-1, 64, 5, 5] 24,640
BatchNorm2d-92 [-1, 64, 5, 5] 128
LinearBottleNeck-93 [-1, 64, 5, 5] 0
Conv2d-94 [-1, 384, 5, 5] 24,960
BatchNorm2d-95 [-1, 384, 5, 5] 768
ReLU6-96 [-1, 384, 5, 5] 0
Conv2d-97 [-1, 384, 5, 5] 3,840
BatchNorm2d-98 [-1, 384, 5, 5] 768
ReLU6-99 [-1, 384, 5, 5] 0
Conv2d-100 [-1, 96, 5, 5] 36,960
BatchNorm2d-101 [-1, 96, 5, 5] 192
LinearBottleNeck-102 [-1, 96, 5, 5] 0
Conv2d-103 [-1, 576, 5, 5] 55,872
BatchNorm2d-104 [-1, 576, 5, 5] 1,152
ReLU6-105 [-1, 576, 5, 5] 0
Conv2d-106 [-1, 576, 5, 5] 5,760
BatchNorm2d-107 [-1, 576, 5, 5] 1,152
ReLU6-108 [-1, 576, 5, 5] 0
Conv2d-109 [-1, 96, 5, 5] 55,392
BatchNorm2d-110 [-1, 96, 5, 5] 192
LinearBottleNeck-111 [-1, 96, 5, 5] 0
Conv2d-112 [-1, 576, 5, 5] 55,872
BatchNorm2d-113 [-1, 576, 5, 5] 1,152
ReLU6-114 [-1, 576, 5, 5] 0
Conv2d-115 [-1, 576, 5, 5] 5,760
BatchNorm2d-116 [-1, 576, 5, 5] 1,152
ReLU6-117 [-1, 576, 5, 5] 0
Conv2d-118 [-1, 96, 5, 5] 55,392
BatchNorm2d-119 [-1, 96, 5, 5] 192
LinearBottleNeck-120 [-1, 96, 5, 5] 0
Conv2d-121 [-1, 576, 5, 5] 55,872
BatchNorm2d-122 [-1, 576, 5, 5] 1,152
ReLU6-123 [-1, 576, 5, 5] 0
Conv2d-124 [-1, 576, 5, 5] 5,760
BatchNorm2d-125 [-1, 576, 5, 5] 1,152
ReLU6-126 [-1, 576, 5, 5] 0
Conv2d-127 [-1, 160, 5, 5] 92,320
BatchNorm2d-128 [-1, 160, 5, 5] 320
LinearBottleNeck-129 [-1, 160, 5, 5] 0
Conv2d-130 [-1, 960, 5, 5] 154,560
BatchNorm2d-131 [-1, 960, 5, 5] 1,920
ReLU6-132 [-1, 960, 5, 5] 0
Conv2d-133 [-1, 960, 5, 5] 9,600
BatchNorm2d-134 [-1, 960, 5, 5] 1,920
ReLU6-135 [-1, 960, 5, 5] 0
Conv2d-136 [-1, 160, 5, 5] 153,760
BatchNorm2d-137 [-1, 160, 5, 5] 320
LinearBottleNeck-138 [-1, 160, 5, 5] 0
Conv2d-139 [-1, 960, 5, 5] 154,560
BatchNorm2d-140 [-1, 960, 5, 5] 1,920
ReLU6-141 [-1, 960, 5, 5] 0
Conv2d-142 [-1, 960, 5, 5] 9,600
BatchNorm2d-143 [-1, 960, 5, 5] 1,920
ReLU6-144 [-1, 960, 5, 5] 0
Conv2d-145 [-1, 160, 5, 5] 153,760
BatchNorm2d-146 [-1, 160, 5, 5] 320
LinearBottleNeck-147 [-1, 160, 5, 5] 0
Conv2d-148 [-1, 960, 5, 5] 154,560
BatchNorm2d-149 [-1, 960, 5, 5] 1,920
ReLU6-150 [-1, 960, 5, 5] 0
Conv2d-151 [-1, 960, 5, 5] 9,600
BatchNorm2d-152 [-1, 960, 5, 5] 1,920
ReLU6-153 [-1, 960, 5, 5] 0
Conv2d-154 [-1, 320, 5, 5] 307,520
BatchNorm2d-155 [-1, 320, 5, 5] 640
LinearBottleNeck-156 [-1, 320, 5, 5] 0
Conv2d-157 [-1, 1280, 5, 5] 410,880
BatchNorm2d-158 [-1, 1280, 5, 5] 2,560
ReLU6-159 [-1, 1280, 5, 5] 0
Conv2d-160 [-1, 100, 1, 1] 128,100
================================================================
Total params: 2,369,380
Trainable params: 2,369,380
Non-trainable params: 0
----------------------------------------------------------------
Input size (MB): 0.01
Forward/backward pass size (MB): 20.18
Params size (MB): 9.04
Estimated Total Size (MB): 29.23
----------------------------------------------------------------
Lightning モジュール
import torch import torch.nn as nn import torch.nn.functional as F from torch.optim.lr_scheduler import OneCycleLR, CyclicLR, ExponentialLR, CosineAnnealingLR, ReduceLROnPlateau from torch.optim.swa_utils import AveragedModel, update_bn import torchvision import pytorch_lightning as pl from pytorch_lightning.callbacks import LearningRateMonitor, GPUStatsMonitor, EarlyStopping from pytorch_lightning.metrics.functional import accuracy #from pl_bolts.datamodules import CIFAR10DataModule #from pl_bolts.transforms.dataset_normalizations import cifar10_normalization
pl.seed_everything(7);
batch_size = 50
train_transforms = torchvision.transforms.Compose([
torchvision.transforms.RandomCrop(32, padding=4),
torchvision.transforms.RandomHorizontalFlip(),
torchvision.transforms.ToTensor(),
cifar100_normalization(),
])
test_transforms = torchvision.transforms.Compose([
torchvision.transforms.ToTensor(),
cifar100_normalization(),
])
cifar100_dm = CIFAR100DataModule(
batch_size=batch_size,
num_workers=8,
train_transforms=train_transforms,
test_transforms=test_transforms,
val_transforms=test_transforms,
)
class LitCifar100(pl.LightningModule):
def __init__(self, lr=0.05, factor=0.8):
super().__init__()
self.save_hyperparameters()
self.model = mobilenetv2()
def forward(self, x):
out = self.model(x)
return F.log_softmax(out, dim=1)
def training_step(self, batch, batch_idx):
x, y = batch
logits = F.log_softmax(self.model(x), dim=1)
loss = F.nll_loss(logits, y)
self.log('train_loss', loss)
return loss
def evaluate(self, batch, stage=None):
x, y = batch
logits = self(x)
loss = F.nll_loss(logits, y)
preds = torch.argmax(logits, dim=1)
acc = accuracy(preds, y)
if stage:
self.log(f'{stage}_loss', loss, prog_bar=True)
self.log(f'{stage}_acc', acc, prog_bar=True)
def validation_step(self, batch, batch_idx):
self.evaluate(batch, 'val')
def test_step(self, batch, batch_idx):
self.evaluate(batch, 'test')
def configure_optimizers(self):
optimizer = torch.optim.SGD(self.parameters(), lr=self.hparams.lr, momentum=0.9, weight_decay=5e-4)
return {
'optimizer': optimizer,
'lr_scheduler': ReduceLROnPlateau(optimizer, 'max', patience=5, factor=self.hparams.factor, verbose=True, threshold=0.0001, threshold_mode='abs', cooldown=1, min_lr=1e-5),
'monitor': 'val_acc'
}
訓練 / 評価
%%time
model = LitCifar100(lr=0.05, factor=0.5)
model.datamodule = cifar100_dm
trainer = pl.Trainer(
gpus=1,
max_epochs=100,
progress_bar_refresh_rate=100,
logger=pl.loggers.TensorBoardLogger('tblogs/', name='mobilenetv2'),
callbacks=[LearningRateMonitor(logging_interval='step')],
)
trainer.fit(model, cifar100_dm)
trainer.test(model, datamodule=cifar100_dm);
| Name | Type | Params
--------------------------------------
0 | model | MobileNetV2 | 2.4 M
--------------------------------------
2.4 M Trainable params
0 Non-trainable params
2.4 M Total params
9.478 Total estimated model params size (MB)
(...)
Epoch 29: reducing learning rate of group 0 to 2.5000e-02.
Epoch 36: reducing learning rate of group 0 to 1.2500e-02.
Epoch 43: reducing learning rate of group 0 to 6.2500e-03.
Epoch 50: reducing learning rate of group 0 to 3.1250e-03.
Epoch 58: reducing learning rate of group 0 to 1.5625e-03.
Epoch 65: reducing learning rate of group 0 to 7.8125e-04.
Epoch 74: reducing learning rate of group 0 to 3.9063e-04.
Epoch 81: reducing learning rate of group 0 to 1.9531e-04.
Epoch 89: reducing learning rate of group 0 to 9.7656e-05.
Epoch 99: reducing learning rate of group 0 to 4.8828e-05.
(...)
--------------------------------------------------------------------------------
DATALOADER:0 TEST RESULTS
{'test_acc': 0.6323999762535095, 'test_loss': 1.4939695596694946}
--------------------------------------------------------------------------------
CPU times: user 1h 14min 38s, sys: 1min 42s, total: 1h 16min 21s
Wall time: 1h 20min 26s
以上