Ecosyste.ms: Awesome
An open API service indexing awesome lists of open source software.
https://github.com/kuan-wang/pytorch-mobilenet-v3
MobileNetV3 in pytorch and ImageNet pretrained models
https://github.com/kuan-wang/pytorch-mobilenet-v3
classification imagenet mobilenet mobilenetv2 mobilenetv3 pytorch
Last synced: about 1 month ago
JSON representation
MobileNetV3 in pytorch and ImageNet pretrained models
- Host: GitHub
- URL: https://github.com/kuan-wang/pytorch-mobilenet-v3
- Owner: kuan-wang
- License: apache-2.0
- Created: 2019-05-08T18:16:57.000Z (over 5 years ago)
- Default Branch: master
- Last Pushed: 2019-06-23T01:31:09.000Z (over 5 years ago)
- Last Synced: 2024-08-01T22:50:11.387Z (4 months ago)
- Topics: classification, imagenet, mobilenet, mobilenetv2, mobilenetv3, pytorch
- Language: Python
- Homepage:
- Size: 50.8 KB
- Stars: 765
- Watchers: 12
- Forks: 187
- Open Issues: 9
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
- awesome-image-classification - unofficial-pytorch : https://github.com/kuan-wang/pytorch-mobilenet-v3
- awesome-image-classification - unofficial-pytorch : https://github.com/kuan-wang/pytorch-mobilenet-v3
- awesome-AutoML-and-Lightweight-Models - kuan-wang/pytorch-mobilenet-v3
README
# A PyTorch implementation of MobileNetV3
This is a PyTorch implementation of MobileNetV3 architecture as described in the paper [Searching for MobileNetV3](https://arxiv.org/pdf/1905.02244.pdf).
Some details may be different from the original paper, welcome to discuss and help me figure it out.
- **[NEW]** The pretrained model of small version mobilenet-v3 is online, accuracy achieves the same as paper.
- **[NEW]** The paper updated on 17 May, so I renew the codes for that, but there still are some bugs.
- **[NEW]** I remove the se before the global avg_pool (the paper may add it in error), and now the model size is close to paper.
## Training & Accuracy
### training setting:
1. number of epochs: 150
2. learning rate schedule: cosine learning rate, initial lr=0.05
3. weight decay: 4e-5
4. remove dropout
5. batch size: 256
### MobileNetV3 large
| | Madds | Parameters | Top1-acc | Pretrained Model |
| ----------- | --------- | ---------- | --------- | ------------------------------------------------------------ |
| Offical 1.0 | 219 M | 5.4 M | 75.2% | - |
| Offical 0.75 | 155 M | 4 M | 73.3% | - |
| Ours 1.0 | 224 M | 5.48 M | 72.8% | - |
| Ours 0.75 | 148 M | 3.91 M | - | - |
### MobileNetV3 small
| | Madds | Parameters | Top1-acc | Pretrained Model |
| ----------- | --------- | ---------- | --------- | ------------------------------------------------------------ |
| Offical 1.0 | 66 M | 2.9 M | 67.4% | - |
| Offical 0.75 | 44 M | 2.4 M | 65.4% | - |
| Ours 1.0 | 63 M | 2.94 M | 67.4% | [[google drive](https://drive.google.com/open?id=1lCsN3kWXAu8C30bQrD2JTZ7S2v4yt23C)] |
| Ours 0.75 | 46 M | 2.38 M | - | - |
## Usage
Pretrained models are still training ...
```python
# pytorch 1.0.1
# large
net_large = mobilenetv3(mode='large')
# small
net_small = mobilenetv3(mode='small')
state_dict = torch.load('mobilenetv3_small_67.4.pth.tar')
net_small.load_state_dict(state_dict)
```
## Data Pre-processing
I used the following code for data pre-processing on ImageNet:
```python
normalize = transforms.Normalize(mean=[0.485, 0.456, 0.406],
std=[0.229, 0.224, 0.225])
input_size = 224
train_loader = torch.utils.data.DataLoader(
datasets.ImageFolder(
traindir, transforms.Compose([
transforms.RandomResizedCrop(input_size),
transforms.RandomHorizontalFlip(),
transforms.ToTensor(),
normalize,
])),
batch_size=batch_size, shuffle=True,
num_workers=n_worker, pin_memory=True)
val_loader = torch.utils.data.DataLoader(
datasets.ImageFolder(valdir, transforms.Compose([
transforms.Resize(int(input_size/0.875)),
transforms.CenterCrop(input_size),
transforms.ToTensor(),
normalize,
])),
batch_size=batch_size, shuffle=False,
num_workers=n_worker, pin_memory=True)
```