Pre-packages Pytorch-Lightning datasets
Project description
LightningData Modules
Pre-packages Pytorch-Lightning datasets.
Installation
To install this library, simply run the following command:
pip install lightingdata_modules
Installing the LightningData Modules should automatically install suitable dependencies.
Customized Dataset Support
LightningData Modules downloads, installs and prepares customized Pytorch Lightning datamodules with just one line of code.
Example from Domain Adaptation (DomainNet):
# import the custom DomainNet module
import lightningdata_modules.domain_adaptation.domainNet_datamodule as domainNet
# initialize the data module with the "real" domain from DomainNet
my_data_module = domainNet.DomainNetDataModule(data_dir="./dataset/", domain="real")
The data module can now be passed to the Pytorch Lighting Trainer instance or be used as a standalone Dataloader:
# check if the dataset already exists in the data_dir, download domain-specific lmdb database from cloud storage otherwise
my_data_module.prepare_data()
# initialize the train and test set according to setup properties
my_data_module.setup()
# create the train DataLoader on-the-fly
train_loader = my_data_module.train_dataloader()
List of Available Datasets
Domain Adaptation | Federated Learning | Meta Learning |
---|---|---|
DomainNet | EMNIST | Omiglot |
Office31 | Mini Imagenet | |
OfficeHome | ||
Digit-Five |
Class Diagram
Ack
Thanks to TorchMeta: https://github.com/tristandeleu/pytorch-meta
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
File details
Details for the file lightningdata-modules-0.1.9.tar.gz
.
File metadata
- Download URL: lightningdata-modules-0.1.9.tar.gz
- Upload date:
- Size: 67.5 kB
- Tags: Source
- Uploaded using Trusted Publishing? No
- Uploaded via: twine/4.0.1 CPython/3.8.12
File hashes
Algorithm | Hash digest | |
---|---|---|
SHA256 | 7aba3152672e9ddbcccc7e59130fe775e41fdb5a6cb629c0517f33e7da9a581e |
|
MD5 | eab40e071185058bf7f4a5e44d58be04 |
|
BLAKE2b-256 | a26900617670585883040d483fc90870f70dcfae1847556e1f00be0a8db135ea |