Skip to content

Fixes issue #2 and #3 #4

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
davidwincent wants to merge 2 commits into
base: master
Choose a base branch
from
Open

Fixes issue #2 and #3 #4

Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
bd8d03b
Fixes Layer Time2Vec has arguments ['self', 'kernel_size', 'periodic_…
davidwincent Jan 5, 2022
2b2fc72
Fixes AttributeError: 'NoneType' object has no attribute 'replace' in…
davidwincent Jan 5, 2022
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
25 changes: 18 additions & 7 deletions Time2Vec/layers.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -3,15 +3,16 @@


class Time2Vec(Layer):
def __init__(self, kernel_size, periodic_activation='sin'):
def __init__(self, kernel_size, periodic_activation='sin', **kwargs):
'''

:param kernel_size: The length of time vector representation.
:param periodic_activation: The periodic activation, sine or cosine, or any future function.
'''
super(Time2Vec, self).__init__(
trainable=True,
name='Time2VecLayer_'+periodic_activation.upper()
name='Time2VecLayer_'+periodic_activation.upper(),
**kwargs
)

self.k = kernel_size
Expand All @@ -22,26 +23,30 @@ def build(self, input_shape):
self.wb = self.add_weight(
shape=(1, 1),
initializer='uniform',
trainable=True
trainable=True,
name='wb_weight'
)

self.bb = self.add_weight(
shape=(1, 1),
initializer='uniform',
trainable=True
trainable=True,
name='bb_weight'
)

# Else needs to pass the periodic activation
self.wa = self.add_weight(
shape=(1, self.k),
initializer='uniform',
trainable=True
trainable=True,
name='wa_weight'
)

self.ba = self.add_weight(
shape=(1, self.k),
initializer='uniform',
trainable=True
trainable=True,
name='ba_weight'
)

super(Time2Vec, self).build(input_shape)
Expand All @@ -63,4 +68,10 @@ def call(self, inputs, **kwargs):
return K.concatenate([bias, wgts], -1)

def compute_output_shape(self, input_shape):
return (input_shape[0], input_shape[1], self.k + 1)
return (input_shape[0], input_shape[1], self.k + 1)

def get_config(self):
config = super(Time2Vec, self).get_config()
config.update({"kernel_size": self.k})
config.update({"periodic_activation": self.p_activation})
return config