steffen/self-replicating-neural-networks
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

upped some setups, data incoming

Browse Source
This commit is contained in:
Thomas Gabor
2019-03-15 04:22:59 +01:00
parent 36a852b3ab
commit 70e19dae1e
3 changed files with 13 additions and 10 deletions
Download Patch File Download Diff File Expand all files Collapse all files

10
code/setups/fixpoint-density.py
View File

@ -31,14 +31,16 @@ def count(counters, net, notable_nets=[]):
if __name__ == '__main__': if __name__ == '__main__':
with Experiment('fixpoint-density') as exp: with Experiment('fixpoint-density') as exp:
exp.trials = 100 #NOTE: settings could/should stay this way
#FFT doesn't work though
exp.trials = 100000
exp.epsilon = 1e-4 exp.epsilon = 1e-4
net_generators = [] net_generators = []
for activation in ['linear', 'sigmoid', 'relu']: for activation in ['linear', 'sigmoid', 'relu']:
# net_generators += [lambda activation=activation: WeightwiseNeuralNetwork(width=2, depth=2).with_keras_params(activation=activation, use_bias=False)] net_generators += [lambda activation=activation: WeightwiseNeuralNetwork(width=2, depth=2).with_keras_params(activation=activation, use_bias=False)]
net_generators += [lambda activation=activation: AggregatingNeuralNetwork(aggregates=4, width=2, depth=2).with_keras_params(activation=activation, use_bias=False)] net_generators += [lambda activation=activation: AggregatingNeuralNetwork(aggregates=4, width=2, depth=2).with_keras_params(activation=activation, use_bias=False)]
# net_generators += [lambda activation=activation: FFTNeuralNetwork(aggregates=4, width=2, depth=2).with_keras_params(activation=activation, use_bias=False)] #net_generators += [lambda activation=activation: FFTNeuralNetwork(aggregates=4, width=2, depth=2).with_keras_params(activation=activation, use_bias=False)]
# net_generators += [lambda activation=activation: RecurrentNeuralNetwork(width=2, depth=2).with_keras_params(activation=activation, use_bias=False)] net_generators += [lambda activation=activation: RecurrentNeuralNetwork(width=2, depth=2).with_keras_params(activation=activation, use_bias=False)]
all_counters = [] all_counters = []
all_notable_nets = [] all_notable_nets = []
all_names = [] all_names = []

5
code/setups/known-fixpoint-variation.py
View File

@ -26,8 +26,9 @@ def generate_fixpoint_weights():
def generate_fixpoint_net(): def generate_fixpoint_net():
# net = WeightwiseNeuralNetwork(width=2, depth=2).with_keras_params(activation='sigmoid') #NOTE: Weightwise only is all we can do right now IMO
net = AggregatingNeuralNetwork(width=2, depth=2).with_keras_params(activation='sigmoid') net = WeightwiseNeuralNetwork(width=2, depth=2).with_keras_params(activation='sigmoid')
# net = AggregatingNeuralNetwork(aggregates=4, width=2, depth=2).with_keras_params(activation='sigmoid') # I don't know if this work for aggregaeting. We don't actually need it, though.
net.set_weights(generate_fixpoint_weights()) net.set_weights(generate_fixpoint_weights())
return net return net

8
code/setups/training-fixpoints.py
View File

@ -33,15 +33,15 @@ def count(counters, net, notable_nets=[]):
if __name__ == '__main__': if __name__ == '__main__':
with Experiment('training_fixpoint') as exp: with Experiment('training_fixpoint') as exp:
exp.trials = 20 exp.trials = 50
exp.run_count = 500 exp.run_count = 1000
exp.epsilon = 1e-4 exp.epsilon = 1e-4
net_generators = [] net_generators = []
for activation in ['linear']: # , 'sigmoid', 'relu']: for activation in ['linear']: # , 'sigmoid', 'relu']:
for use_bias in [False]: for use_bias in [False]:
# net_generators += [lambda activation=activation, use_bias=use_bias: WeightwiseNeuralNetwork(width=2, depth=2).with_keras_params(activation=activation, use_bias=use_bias)] net_generators += [lambda activation=activation, use_bias=use_bias: WeightwiseNeuralNetwork(width=2, depth=2).with_keras_params(activation=activation, use_bias=use_bias)]
net_generators += [lambda activation=activation, use_bias=use_bias: AggregatingNeuralNetwork(aggregates=4, width=2, depth=2).with_keras_params(activation=activation, use_bias=use_bias)] net_generators += [lambda activation=activation, use_bias=use_bias: AggregatingNeuralNetwork(aggregates=4, width=2, depth=2).with_keras_params(activation=activation, use_bias=use_bias)]
# net_generators += [lambda activation=activation, use_bias=use_bias: RecurrentNeuralNetwork(width=2, depth=2).with_keras_params(activation=activation, use_bias=use_bias)] net_generators += [lambda activation=activation, use_bias=use_bias: RecurrentNeuralNetwork(width=2, depth=2).with_keras_params(activation=activation, use_bias=use_bias)]
all_counters = [] all_counters = []
all_notable_nets = [] all_notable_nets = []
all_names = [] all_names = []