rewrite some comments

This commit is contained in:
Connor Olding 2019-02-03 14:30:58 +01:00
parent 94f27d6f2a
commit b8c40d2e2f

View File

@ -122,7 +122,7 @@ class RMSprop(Optimizer):
class RMSpropCentered(Optimizer):
# referenced TensorFlow/PyTorch.
# paper: https://arxiv.org/pdf/1308.0850v5.pdf
# paper: https://arxiv.org/abs/1308.0850v5
def __init__(self, lr=1e-4, aleph=0.95, momentum=0.9, eps=1e-8):
self.aleph = _f(aleph)
@ -532,11 +532,11 @@ class PowerSign(Optimizer):
class Neumann(Optimizer):
# paper: https://arxiv.org/abs/1712.03298
# NOTE: this implementation is missing resetting as described in the paper.
# NOTE: this implementation omits resetting as described in the paper.
# resetting is totally disabled for now.
# NOTE: this implementation does not use vanilla SGD for its first epochs.
# you should do this yourself if you need it.
# it seems like using a Learner like SineCLR makes this unnecessary.
# you can do this yourself if you really want to.
# it seems to be enough to use a slow-starting Learner like SineCLR.
def __init__(self, lr=0.01):
self.alpha = _f(1e-7) # cubic.