Ecosyste.ms: Awesome
An open API service indexing awesome lists of open source software.
https://github.com/flyyufelix/DenseNet-Keras
DenseNet Implementation in Keras with ImageNet Pretrained Models
https://github.com/flyyufelix/DenseNet-Keras
Last synced: 3 months ago
JSON representation
DenseNet Implementation in Keras with ImageNet Pretrained Models
- Host: GitHub
- URL: https://github.com/flyyufelix/DenseNet-Keras
- Owner: flyyufelix
- License: mit
- Created: 2017-05-12T08:39:09.000Z (over 7 years ago)
- Default Branch: master
- Last Pushed: 2019-10-12T18:53:41.000Z (over 5 years ago)
- Last Synced: 2024-08-01T22:49:56.787Z (6 months ago)
- Language: Python
- Homepage:
- Size: 166 KB
- Stars: 566
- Watchers: 29
- Forks: 263
- Open Issues: 15
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
- awesome-image-classification - unofficial-keras : https://github.com/flyyufelix/DenseNet-Keras
- awesome-image-classification - unofficial-keras : https://github.com/flyyufelix/DenseNet-Keras
README
# DenseNet-Keras with ImageNet Pretrained Models
This is an [Keras](https://keras.io/) implementation of DenseNet with [ImageNet](http://www.image-net.org/) pretrained weights. The weights are converted from [Caffe Models](https://github.com/shicai/DenseNet-Caffe). The implementation supports both [Theano](http://deeplearning.net/software/theano/) and [TensorFlow](https://www.tensorflow.org/) backends.
To know more about how DenseNet works, please refer to the [original paper](https://arxiv.org/abs/1608.06993)
```
Densely Connected Convolutional Networks
Gao Huang, Zhuang Liu, Kilian Q. Weinberger, Laurens van der Maaten
arXiv:1608.06993
```
## Pretrained DenseNet Models on ImageNet
The top-1/5 accuracy rates by using single center crop (crop size: 224x224, image size: 256xN)
Network|Top-1|Top-5|Theano|Tensorflow
:---:|:---:|:---:|:---:|:---:
DenseNet 121 (k=32)| 74.91| 92.19| [model (32 MB)](https://drive.google.com/open?id=0Byy2AcGyEVxfMlRYb3YzV210VzQ)| [model (32 MB)](https://drive.google.com/open?id=0Byy2AcGyEVxfSTA4SHJVOHNuTXc)
DenseNet 169 (k=32)| 76.09| 93.14| [model (56 MB)](https://drive.google.com/open?id=0Byy2AcGyEVxfN0d3T1F1MXg0NlU)| [model (56 MB)](https://drive.google.com/open?id=0Byy2AcGyEVxfSEc5UC1ROUFJdmM)
DenseNet 161 (k=48)| 77.64| 93.79| [model (112 MB)](https://drive.google.com/open?id=0Byy2AcGyEVxfVnlCMlBGTDR3RGs)| [model (112 MB)](https://drive.google.com/open?id=0Byy2AcGyEVxfUDZwVjU2cFNidTA)
## Usage
First, download the above pretrained weights to the `imagenet_models` folder.
Run `test_inference.py` for an example of how to use the pretrained model to make inference.
```
python test_inference.py
```
## Fine-tuning
Check [this](https://github.com/flyyufelix/cnn_finetune) out to see example of fine-tuning DenseNet with your own dataset.
## Requirements
* Keras ~~1.2.2~~ 2.0.5
* Theano 0.8.2 or TensorFlow ~~0.12.0~~ 1.2.1
## Updates
* Keras 2.0.5 and TensorFlow 1.2.1 are supported