keras

Version:: 2.3.1
Category:: ai
Cluster:: Loki

Author / Distributor

Description

Keras is a high-level neural networks API, written in Python and capable of running on top of TensorFlow, CNTK, or Theano. It allows fast experimentation and supports both convolutional and recurrent networks, as well as combinations of the two.

Key features:

User-friendly and modular
Seamless integration with TensorFlow backend
GPU acceleration via CUDA
Supports multi-GPU and distributed training
Suited for both research and production environments

Documentation

$ python -m keras --help

usage: keras [-h] {train,test,predict} ...

Keras command-line interface is minimal by default.

Most workflows are run as Python scripts importing keras models:

    from keras.models import Sequential
    from keras.layers import Dense

    model = Sequential()
    model.add(Dense(64, activation='relu'))
    model.add(Dense(10, activation='softmax'))

    model.compile(optimizer='adam',
                  loss='categorical_crossentropy',
                  metrics=['accuracy'])

    model.fit(x_train, y_train, epochs=5, batch_size=32)

Full documentation: https://keras.io/

Examples/Usage

Load the Keras module:

$ module load ai/keras-py37-cuda10.2-gcc/2.3.1

Unload the module:

$ module unload ai/keras-py37-cuda10.2-gcc/2.3.1

Use Keras in a Python script:

import keras
from keras.models import Sequential
from keras.layers import Dense

model = Sequential()
model.add(Dense(128, activation='relu', input_shape=(784,)))
model.add(Dense(10, activation='softmax'))

model.compile(optimizer='adam',
              loss='categorical_crossentropy',
              metrics=['accuracy'])

print(model.summary())

Installation

Source code is obtained from Keras GitHub