Refactor:

Step 1 - Introduction of Weight object for global weight operations
Step2 - Cleanup
Step 3 - Redone WEightwise network updates in clean numpy code
This commit is contained in:
Si11ium
2019-06-06 21:57:22 +02:00
parent f3987cdbb5
commit 50f7f84084
14 changed files with 193 additions and 865 deletions

View File

@ -3,16 +3,18 @@ import os
# Concat top Level dir to system environmental variables
sys.path += os.path.join('..', '.')
from util import *
from experiment import *
from network import *
import keras.backend
import tensorflow.python.keras.backend as K
def generate_counters():
return {'divergent': 0, 'fix_zero': 0, 'fix_other': 0, 'fix_sec': 0, 'other': 0}
def count(counters, net, notable_nets=[]):
def count(counters, net, notable_nets=None):
notable_nets = notable_nets or []
if net.is_diverged():
counters['divergent'] += 1
elif net.is_fixpoint():
@ -52,7 +54,7 @@ if __name__ == '__main__':
net = ParticleDecorator(net)
name = str(net.__class__.__name__) + " activiation='" + str(net.get_keras_params().get('activation')) + "' use_bias='" + str(net.get_keras_params().get('use_bias')) + "'"
count(counters, net, notable_nets)
keras.backend.clear_session()
K.clear_session()
all_counters += [counters]
# all_notable_nets += [notable_nets]
all_names += [name]