coyote is a deep learning library developed for practical applications such as NLP, robotics.
from coyote.Layer import *
from coyote.Model import *
# make neural network
layers=[]
layers.append(FullyConnect(n_in=28 * 28, n_out=100, activation='tanh'))
layers.append(FullyConnect(n_in=100, n_out=50, activation='tanh'))
layers.append(LogisticRegression(n_in=50, n_out=10))
# compile model and train
model = Model(layers)
model.fit(train_set_x, train_set_y,validation_data=[valid_set_x,valid_set_y])
###Philosophy of Coyote
- simple: we provide intuitive and easy to understand interface for everyone from a beginner to a expert.
- flexible input and output: you can directly connect your data source such as images or text to Coyote.
- task oriented: we will implements task oriented methods such as deep q-learning, sequence to sequence learning.