https://github.com/pedro-r-marques/keras-opt

Keras Scipy optimize interface
https://github.com/pedro-r-marques/keras-opt

keras optimization scipy

Last synced: 3 days ago
JSON representation

Keras Scipy optimize interface

• Host: GitHub
• URL: https://github.com/pedro-r-marques/keras-opt
• Owner: pedro-r-marques
• License: apache-2.0
• Created: 2019-03-20T14:38:41.000Z (over 5 years ago)
• Default Branch: master
• Last Pushed: 2020-12-18T17:28:29.000Z (almost 4 years ago)
• Last Synced: 2024-01-30T00:41:23.929Z (10 months ago)
• Topics: keras, optimization, scipy
• Language: Python
• Size: 42 KB
• Stars: 15
• Watchers: 2
• Forks: 2
• Open Issues: 2
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

README

        
# keras-opt

Keras Scipy optimize interface

Interface between keras optimizers and scipy.optimize. It is used to run full

batch optimization rather than mini-batch stochastic gradient descent. It

is applicable to factorization of very sparse matrices where stochastic

gradient descent is not able to converge.

Example usage:

```python

#%%

# Model definition (linear regression)

from keras_opt import scipy_optimizer

from tensorflow.keras.models import Sequential

from tensorflow.keras.layers import Dense, InputLayer

model = Sequential()

model.add(InputLayer(input_shape=(4,)))

model.add(Dense(1, use_bias=False))

model.compile(loss='mse')

#%%

# Generate test data

import numpy as np

np.random.seed(42)

X = np.random.uniform(size=40).reshape(10, 4)

y = np.dot(X, np.array([1, 2, 3, 4])[:, np.newaxis])

#%%

# Use scipy.optimize to minimize the cost

model.train_function = scipy_optimizer.make_train_function(

            model, maxiter=20)

history = model.fit(X, y)

#%%

# Show weights.

model.trainable_weights

```