Usage
ConvModel(n_filters = 12, filter_shape = c(3, 3), activ_func = "logistic_regression", transform_method = "tf.expand_dims", pool_method = "tf.reduce_max", reduction_indices = c(1, 2), shape = c(-1, 12), ...)
Arguments
n_filters
Number of filters
filter_shape
A vector of integers representing the shape of filters
activ_func
The activation function, e.g. logistic_regression, linear_regression
transform_method
The transformation method used, a list of methods can be found on TensorFlow API doc
pool_method
Pooling method. A list can be found on API doc
reduction_indices
A vector of reduction indices used in pooling
shape
The shape parameter passed to tf.reshape. See API doc.
...
Additional parameters can be passed, such as strides, see a list here: https://github.com/google/skflow/blob/master/skflow/ops/conv_ops.py