Torch implementation of Soft-DTW, supports CUDA devices.
Project description
pysdtw
Torch implementation of the Soft-DTW algorithm, supports both cpu and CUDA hardware.
Note: This repository started as a fork from this project.
Installation
This package is available on pypi and depends on pytorch
and numba
.
Install with:
pip install pysdtw
Usage
import pysdtw
# the input data includes a batch dimension
X = torch.rand((10, 5, 7), requires_grad=True)
Y = torch.rand((10, 9, 7))
# optionally choose a pairwise distance function
fun = pysdtw.distance.pairwise_l2_squared
# create the SoftDTW distance function
sdtw = pysdtw.SoftDTW(gamma=1.0, dist_func=fun, use_cuda=False)
# soft-DTW discrepancy, approaches DTW as gamma -> 0
res = sdtw(X, Y)
# define a loss, which gradient can be backpropagated
loss = res.sum()
loss.backward()
# X.grad now contains the gradient with respect to the loss
Project details
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distribution
pysdtw-0.0.5.tar.gz
(6.1 kB
view details)
Built Distribution
File details
Details for the file pysdtw-0.0.5.tar.gz
.
File metadata
- Download URL: pysdtw-0.0.5.tar.gz
- Upload date:
- Size: 6.1 kB
- Tags: Source
- Uploaded using Trusted Publishing? No
- Uploaded via: twine/4.0.1 CPython/3.9.13
File hashes
Algorithm | Hash digest | |
---|---|---|
SHA256 | a1e9a25c250f0da3a886dc7d779e9f3ab8b8cbbdba607427a9fef146b25ae4e8 |
|
MD5 | 731dee534fd5930b196f6c90140bc52c |
|
BLAKE2b-256 | 517ce201ffa0144916bd2c26a7e5a2d2aeb6f1ecadd89502bdded9d6deb830c1 |
File details
Details for the file pysdtw-0.0.5-py3-none-any.whl
.
File metadata
- Download URL: pysdtw-0.0.5-py3-none-any.whl
- Upload date:
- Size: 7.1 kB
- Tags: Python 3
- Uploaded using Trusted Publishing? No
- Uploaded via: twine/4.0.1 CPython/3.9.13
File hashes
Algorithm | Hash digest | |
---|---|---|
SHA256 | 1a46d46fb00408a300a2dbe006ea957d3efef20502cc98d24c5b22e3ee867653 |
|
MD5 | 4f1ef667eb9e3b2894ff93316d0c236b |
|
BLAKE2b-256 | ac0062c234127146cc4695fe2e28fb827c59b4f2d417deb09b9a0c21eb99781f |