notwa/optim
1
0
Fork 0
Code Issues Pull requests Releases Wiki Activity

fix code i forgot to test, plus some tweaks

Browse source
This commit is contained in:
Connor Olding 2017-07-11 11:36:11 +00:00
parent 7bd5518650
commit 112e263056
2 changed files with 3 additions and 4 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

5
onn.py
View file

@ -1014,11 +1014,10 @@ def run(program, args=None):
#optim = 'ftml', #optim = 'ftml',
#optim_decay1 = 2, #optim_decay1 = 2,
#optim_decay2 = 100, #optim_decay2 = 100,
#nesterov = False,
optim = 'adam', # note: most features only implemented for Adam optim = 'adam', # note: most features only implemented for Adam
optim_decay1 = 24, # first momentum given in epochs (optional) optim_decay1 = 24, # first momentum given in epochs (optional)
optim_decay2 = 100, # second momentum given in epochs (optional) optim_decay2 = 100, # second momentum given in epochs (optional)
nesterov = True, nesterov = True, # not available for all optimizers.
batch_size = 64, batch_size = 64,
# learning parameters # learning parameters
@ -1092,7 +1091,7 @@ def run(program, args=None):
def measure_error(): def measure_error():
def print_error(name, inputs, outputs, comparison=None): def print_error(name, inputs, outputs, comparison=None):
predicted = model.forward(inputs) predicted = model.forward(inputs)
err = ritual.measure(predicted, outputs) err = ritual.mloss.forward(predicted, outputs)
if config.log10_loss: if config.log10_loss:
print(name, "{:12.6e}".format(err)) print(name, "{:12.6e}".format(err))
if comparison: if comparison:

2
onn_core.py
View file

@ -1025,7 +1025,7 @@ class Ritual: # i'm just making up names at this point.
avg_loss = self.cumsum_loss / _f(batch_count) avg_loss = self.cumsum_loss / _f(batch_count)
return avg_loss, avg_mloss, self.losses, self.mlosses return avg_loss, avg_mloss, self.losses, self.mlosses
elif return_losses: elif return_losses:
return avg_mloss, mlosses return avg_mloss, self.mlosses
return avg_mloss return avg_mloss
def test_batched(self, inputs, outputs, *args, **kwargs): def test_batched(self, inputs, outputs, *args, **kwargs):