Split¶
Description¶
This module performs the operation of splitting the tensor values along a given axis.
Initializing¶
def __init__(self, axis, sections, name=None):
Parameters
| Parameter | Allowed types | Description | Default |
|---|---|---|---|
| axis | int | Axis on which the operation is performed | - |
| sections | tuple | nto what parts to break the specified axis. The number of output tensors is equal to the length of a given tuple | - |
| name | str | Layer name | None |
Explanations
-
Examples¶
Necessary imports.
import numpy as np
from PuzzleLib.Backend import gpuarray
from PuzzleLib.Modules import Split
Info
gpuarray is required to properly place the tensor in the GPU
batchsize, groups, size = 5, 3, 6
data = gpuarray.to_gpu(np.random.randn(batchsize, groups, size).astype(np.float32))
split = Split(axis=2, sections=(3, 2, 1))
outdata = split(data)
for outd in outdata:
... print(outd.shape)
(5, 3, 3)
(5, 3, 2)
(5, 3, 1)