69 lines
2.4 KiB
Python
69 lines
2.4 KiB
Python
"""
|
|
forward propagator
|
|
|
|
Usage:
|
|
forward_propagator.py -i INPUTS -w WEIGHTS [--a=<ACTUAL>] [--act=<ACTIVATOR>]
|
|
|
|
Options:
|
|
-h, --help Show this help message
|
|
-i INPUTS Comma separated list of inputs
|
|
-w WEIGHTS Comma separated list of weights
|
|
--a=<ACTUAL> The actual value
|
|
--act=<ACTIVATOR> The activator function
|
|
"""
|
|
import docopt
|
|
|
|
activators = {
|
|
'identity': lambda x: x,
|
|
'relu': lambda x: max(0, x),
|
|
'binstep': lambda x: 0 if x < 0 else 1,
|
|
'logistic': lambda x: 1 / (1 + math.exp(-x)),
|
|
'softsign': lambda x: x / (1 + abs(x)),
|
|
'leakyrelu': lambda x: 0.01 * x if x < 0 else x
|
|
}
|
|
|
|
def forward_propagator(inputs, weights, actual=0, activator='relu'):
|
|
if ((len(weights) - len(inputs)) % pow(len(inputs), 2)) != 0:
|
|
raise Exception("Incorrect number of weights provided.")
|
|
|
|
try:
|
|
activator = activators[activator]
|
|
except KeyError:
|
|
activator = activators['identity']
|
|
|
|
dot = lambda X, Y: sum(map(lambda x, y: x * y, X, Y))
|
|
|
|
weights_per_epoch = pow(len(inputs), 2)
|
|
|
|
num_epochs = (len(weights) - len(inputs)) / weights_per_epoch
|
|
node_output_weights = weights[-len(inputs):]
|
|
|
|
print("weights per epoch={}, number of epochs={}".format(weights_per_epoch, num_epochs))
|
|
node_inputs = inputs
|
|
|
|
for epoch in range(0, num_epochs):
|
|
epoch_weights = weights[epoch * weights_per_epoch:(epoch + 1) * weights_per_epoch]
|
|
node_input_weights = [epoch_weights[i:i + len(inputs)] for i in range(0, len(epoch_weights), len(inputs))]
|
|
print("epoch: {}".format(epoch))
|
|
print("node inputs: {}".format(node_inputs))
|
|
print("epoch weights: {}".format(epoch_weights))
|
|
print("node input weights: {}".format(node_input_weights))
|
|
|
|
sums = [activator(dot(node_weights, node_inputs)) for node_weights in node_input_weights]
|
|
print("sums: {}".format(sums))
|
|
|
|
print("")
|
|
|
|
node_inputs = sums
|
|
ret = dot(node_output_weights, node_inputs)
|
|
|
|
return ret - actual
|
|
|
|
if __name__ == "__main__":
|
|
args = docopt.docopt(__doc__)
|
|
args['-i'] = [float(x) for x in args['-i'].split(',')]
|
|
args['-w'] = [float(x) for x in args['-w'].split(',')]
|
|
actual = float(args['--a']) if args['--a'] is not None else 0
|
|
activator = args['--act'] if args['--act'] is not None else 'identity'
|
|
print(forward_propagator(inputs=args['-i'], weights=args['-w'], actual=actual, activator=activator))
|