TorchLikelihoods: User-friendly handling of likelihoods in Pytorch
Project description
TorchLikelihoods
A library for handling likelihoods in PyTorch for any type of data
Installation
Run the following to install
pip install torchlikelihoods
Usage
from torchlikelihoods import NormalLikelihood
import torch
num_samples, num_feats = 100, 5
normal_data = torch.randn((num_samples, num_feats))
lik = NormalLikelihood(domain_size=num_feats)
scaler = get_scaler
print(f"Domain size: {lik.domain_size()}")
print(f"Params size: {lik.params_size()}")
Do you want to get involved in the development?
pip install -e .[dev]
Testing
To run the tests:
make test
pytest
Create source distribution
python setup.py sdist
tar tzf dist/torchlikelihoods-0.0.1.tar.gz
To include all the source code
pip install check-manifest
check-manifest --create
Publish it!
python setup.py bdist_wheel sdist
Project details
Release history Release notifications | RSS feed
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distribution
torchlikelihoods-0.0.1.tar.gz
(6.7 kB
view hashes)
Built Distribution
Close
Hashes for torchlikelihoods-0.0.1-py3-none-any.whl
Algorithm | Hash digest | |
---|---|---|
SHA256 | edb69e69e13a240d32f325b2e498b26d98712b0685490b1b8c2eaf44c0400e01 |
|
MD5 | 09d4f3bdd60e5791ebab0f1eb7b2af31 |
|
BLAKE2b-256 | 3532c91cde811737ccd2a35201f9b78a2b64da4b783c0dcbb2afd5b35715aeed |