Ecosyste.ms: Awesome
An open API service indexing awesome lists of open source software.
https://github.com/asmith26/wide_resnets_keras
Keras implementation + pretrained weights for "Wide Residual Networks"
https://github.com/asmith26/wide_resnets_keras
Last synced: 4 days ago
JSON representation
Keras implementation + pretrained weights for "Wide Residual Networks"
- Host: GitHub
- URL: https://github.com/asmith26/wide_resnets_keras
- Owner: asmith26
- License: apache-2.0
- Created: 2016-06-01T16:44:46.000Z (over 8 years ago)
- Default Branch: main
- Last Pushed: 2024-01-18T22:48:51.000Z (10 months ago)
- Last Synced: 2024-08-01T22:49:53.805Z (3 months ago)
- Language: Python
- Homepage: http://arxiv.org/abs/1605.07146v1
- Size: 33 MB
- Stars: 138
- Watchers: 6
- Forks: 47
- Open Issues: 6
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
- awesome-image-classification - unofficial-keras : https://github.com/asmith26/wide_resnets_keras
- awesome-image-classification - unofficial-keras : https://github.com/asmith26/wide_resnets_keras
README
# Keras implementation of "Wide Residual Networks"
This repo contains the code to run Wide Residual Networks using Keras.
- Paper (v1): http://arxiv.org/abs/1605.07146v1 (the authors have since published a v2 of the paper, which introduces slightly different preprocessing and improves the accuracy a little).
- Original code: https://github.com/szagoruyko/wide-residual-networks
## Dependencies:
- `pip install -r requirements.txt`
- To plot the architecture of the model used (like the plot of the WRN-16-2 architecture plotted [below](#example-plot)), you need to install `pydot` and `graphviz`. I recommend installing with `conda install -c conda-forge python-graphviz`:
## Training Details:
Run the default configuration (i.e. best configuration for CIFAR10 from original paper/code, WRN-28-10 without dropout) with:
```
$ python main.py
```
There are three configuration sections at the top of `main.py`:
- [DATA CONFIGURATION](https://github.com/asmith26/wide_resnets_keras/blob/master/main.py#L34-48): Containing data details.
- [NETWORK/TRAINING CONFIGURATION](https://github.com/asmith26/wide_resnets_keras/blob/master/main.py#L50-87): Includes the main parameters the authors experimented with.
- [OUTPUT CONFIGURATION](https://github.com/asmith26/wide_resnets_keras/blob/master/main.py#L89-97): Defines paths regarding where to save model/checkpoint weights and plots.
## Results and Trained models:
- ***WRN-28-10 no dropout***:
- Using these values in **main.py**, I obtained a **test loss = 0.31** and **test accuracy = 0.93**. This test error (i.e. 1 - 0.93 = **7%**) is a little higher than the reported result (Table 4 states the same model obtains a test error of *4.97%*); see the note below for a likely explanation.
- You can find the trained weights for this model at **models/WRN-28-10.h5**, whilst **[models/test.py](https://github.com/asmith26/wide_resnets_keras/blob/master/models/test.py)** provides an example of running these weights against the test set.
**Note:** I have not followed the exact same preprocessing and data augmentation steps used in the paper, in particular:
- "global *contrast* normalization", and
- "random crops from image padded by 4 pixels on each side, filling missing pixels with reflections of original image", which appears to be implemented in [this file](https://github.com/szagoruyko/wide-residual-networks/blob/8b166cc15fa8a598490ce0ae66365bf165dffb75/augmentation.lua).
Ideally, we will add such methods directly to the [Keras image preprocessing script](https://github.com/fchollet/keras/blob/master/keras/preprocessing/image.py).
## WRN-16-2 Architecture
