ShuangXieIrene / Mobilenet V2
Repository for "Inverted Residuals and Linear Bottlenecks: Mobile Networks for Classification, Detection and Segmentation".
Stars: ✭ 73
Programming Languages
python
139335 projects - #7 most used programming language
Projects that are alternatives of or similar to Mobilenet V2
MobileNet-SSD-windows
No description or website provided.
Stars: ✭ 91 (+24.66%)
Mutual labels: mobilenet, mobilenetv2
Ssds pytorch
Multiple basenet MobileNet v1,v2, ResNet combined with SSD detection method and it's variants such as RFB, FSSD etc.
Stars: ✭ 71 (-2.74%)
Mutual labels: mobilenet, mobilenetv2
MobileNet V2 Keras
No description or website provided.
Stars: ✭ 29 (-60.27%)
Mutual labels: mobilenet, mobilenetv2
caffe-mobilenet v2
caffe based mobilenet v2 deploy
Stars: ✭ 29 (-60.27%)
Mutual labels: mobilenet, mobilenetv2
Tensornets
High level network definitions with pre-trained weights in TensorFlow
Stars: ✭ 982 (+1245.21%)
Mutual labels: mobilenet, mobilenetv2
Object Localization
Object localization in images using simple CNNs and Keras
Stars: ✭ 130 (+78.08%)
Mutual labels: mobilenet, mobilenetv2
mobilenet-v2-tensorflow
No description or website provided.
Stars: ✭ 66 (-9.59%)
Mutual labels: mobilenet, mobilenetv2
Mobilenet Caffe
Caffe Implementation of Google's MobileNets (v1 and v2)
Stars: ✭ 1,217 (+1567.12%)
Mutual labels: mobilenet, mobilenetv2
Mobilenetv2 Ssdlite
Caffe implementation of SSD and SSDLite detection on MobileNetv2, converted from tensorflow.
Stars: ✭ 435 (+495.89%)
Mutual labels: mobilenet, mobilenetv2
yolo3 tensorflow
yolo3 implement by tensorflow， including mobilenet_v1, mobilenet_v2
Stars: ✭ 48 (-34.25%)
Mutual labels: mobilenet, mobilenetv2
LightNet
LightNet: Light-weight Networks for Semantic Image Segmentation (Cityscapes and Mapillary Vistas Dataset)
Stars: ✭ 710 (+872.6%)
Mutual labels: mobilenet, mobilenetv2
Lightnet
LightNet: Light-weight Networks for Semantic Image Segmentation (Cityscapes and Mapillary Vistas Dataset)
Stars: ✭ 698 (+856.16%)
Mutual labels: mobilenet, mobilenetv2
Pytorch Mobilenet V3
MobileNetV3 in pytorch and ImageNet pretrained models
Stars: ✭ 616 (+743.84%)
Mutual labels: mobilenet, mobilenetv2
Seg Mentor
TFslim based semantic segmentation models, modular&extensible boutique design
Stars: ✭ 43 (-41.1%)
Mutual labels: mobilenet, slim
Mcp Panthor
A thin PHP microframework built on Slim and Symfony
Stars: ✭ 11 (-84.93%)
Mutual labels: slim
Jacinto Ai Devkit
Training & Quantization of embedded friendly Deep Learning / Machine Learning / Computer Vision models
Stars: ✭ 49 (-32.88%)
Mutual labels: mobilenet
Classification models
Classification models trained on ImageNet. Keras.
Stars: ✭ 938 (+1184.93%)
Mutual labels: mobilenet
Node Image Classification
A simple example on Image Classification in Node.js with TensorFlow.js
Stars: ✭ 23 (-68.49%)
Mutual labels: mobilenet
Elliot
Comprehensive and Rigorous Framework for Reproducible Recommender Systems Evaluation
Stars: ✭ 49 (-32.88%)
Mutual labels: slim
MobilenetV2.tensorflow
Repository for "Inverted Residuals and Linear Bottlenecks: Mobile Networks for Classification, Detection and Segmentation".
Pre-request libraries
- python2/3
- tensorflow
- tensorflow-slim (optional)
Training & Accuracy
The model was trained with SGD+Momentum optimizer.
| Optimizer | iter | Top1 | Top5 | Pre-trained Model |
|---|---|---|---|---|
| SGD | 1million | 69.7% | 89.2% | google drive |
Usage
The MobilenetV2 is implemented by two versions:
- By tensorflow-slim. This version can be copied under tensorflow-slim and use the API which is provided by slim to train and test the model.
- By tensorflow. This version is used to achieved higher speed with NCHW and half-percision.
Network Architecture
Please notice there is a mistake in the paper (after the fourth bottleneck layer, the output size should be × 64 instead of
× 64 )
| Input | Operator | t | c | n | s |
|---|---|---|---|---|---|
|
|
conv2d | - | 32 | 1 | 2 |
|
|
bottleneck | 1 | 16 | 1 | 1 |
|
|
bottleneck | 6 | 24 | 2 | 2 |
|
|
bottleneck | 6 | 32 | 3 | 2 |
|
|
bottleneck | 6 | 64 | 4 | 2 |
|
|
bottleneck | 6 | 96 | 3 | 1 |
|
|
bottleneck | 6 | 160 | 3 | 2 |
|
|
bottleneck | 6 | 320 | 1 | 1 |
|
|
conv2d 1x1 | - | 1280 | 1 | 1 |
|
|
avgpool 7x7 | - | - | 1 | - |
| 1 × 1 × k | conv2d 1x1 | - | k |
Each line describes a sequence of 1 or more identical (modulo stride) layers, repeated n times. All layers in the same sequence have the same number c of output channels. The first layer of each sequence has a stride s and all others use stride 1. All spatial convolutions use 3 × 3 kernels. The expansion factor t is always applied to the input size.
Linear Bottlenecks architecture
The bottleneck blocks use shortcut directly between the bottlenecks to improve the ability of gradient to propagate across multipler layers.
- Because ReLU is capable of preserving information only if the input manifold lies in a low-dimensional subspace of the input space, the input and output data of linear bottlenecks is in low-dimension space.
- Stride (s) in the depthwise convolutions is for down sampling.
- Expansion ratio (t) is greater than 1 for projecting the low dimensioal data to high dimension.
TODO
- [x] Pre-trained model
- [ ] Time test in 1080Ti
Note that the project description data, including the texts, logos, images, and/or trademarks,
for each open source project belongs to its rightful owner.
If you wish to add or remove any projects, please contact us at [email protected].