Keraflow
Deep Learning for Python.
|
Built-in initialization functions. More...
Functions | |
def | uniform |
Unifrom initialization. | |
def | normal |
Normal initialization. | |
def | lecun_uniform |
Uniform initialization scaled by the square root of the number of inputs (LeCun 98). More... | |
def | glorot_normal |
Gaussian initialization scaled by fan_in + fan_out. More... | |
def | glorot_uniform |
Uniform initialization scaled by fan_in + fan_out. | |
def | he_normal |
Gaussian initialization scaled by fan_in. More... | |
def | he_uniform |
he_uniform initialization. | |
def | orthogonal |
Use with square 2D layers (shape[0] == shape[1] ). More... | |
def | identity |
Use with square 2D layers (shape[0] == shape[1] ). | |
def | zero |
All zeros initialization. | |
def | one |
All ones initialization. | |
Built-in initialization functions.
def keraflow.initializations.glorot_normal | ( | shape, | |
name = None |
|||
) |
Gaussian initialization scaled by fan_in + fan_out.
def keraflow.initializations.he_normal | ( | shape, | |
name = None |
|||
) |
Gaussian initialization scaled by fan_in.
def keraflow.initializations.lecun_uniform | ( | shape, | |
name = None |
|||
) |
Uniform initialization scaled by the square root of the number of inputs (LeCun 98).
def keraflow.initializations.orthogonal | ( | shape, | |
scale = 1.1 , |
|||
name = None |
|||
) |
Use with square 2D layers (shape[0] == shape[1]
).