/keras-squeezenet

SqueezeNet implementation with Keras Framework

Primary LanguagePythonMIT LicenseMIT

keras-squeezenet Build Status

SqueezeNet v1.1 Implementation using Keras Functional Framework 2.0

This network model has AlexNet accuracy with small footprint (5.1 MB) Pretrained models are converted from original Caffe network.

# Most Recent One
pip install git+https://github.com/rcmalli/keras-squeezenet.git
# Release Version
pip install keras_squeezenet

News

  • Project is now up-to-date with the new Keras version (2.0).

  • Old Implementation is still available at 'keras1' branch but not updated.

Library Versions

  • Keras v2.1.1
  • Tensorflow v1.4

Example Usage

import numpy as np
from keras_squeezenet import SqueezeNet
from keras.applications.imagenet_utils import preprocess_input, decode_predictions
from keras.preprocessing import image

model = SqueezeNet()

img = image.load_img('../images/cat.jpeg', target_size=(227, 227))
x = image.img_to_array(img)
x = np.expand_dims(x, axis=0)
x = preprocess_input(x)

preds = model.predict(x)
print('Predicted:', decode_predictions(preds))

References

  1. Keras Framework

  2. SqueezeNet Official Github Repo

  3. SqueezeNet Paper

Licence

MIT License

Note: If you find this project useful, please include reference link in your work.