Easy-to-use PyTorch library for cross-domain learning, few-shot learning and meta-learning
Project description
TorchCross
Easy-to-use PyTorch library for cross-domain learning, few-shot learning and meta-learning.
What is TorchCross?
TorchCross is a PyTorch library for cross-domain learning, few-shot learning and meta-learning. It provides convenient utilities for creating cross-domain learning or few-shot learning experiments.
Package Overview
torchcross
: The main package, containing the core functionality of the library.torchcross.data
: Contains theCrossDomainDataset
andFewShotDataset
classes, which wrapTaskSource
instances to produce batches for cross-domain learning or tasks for few-shot learning experiments.torchcross.data.task
: Contains theTask
andTaskDescription
classes, which represent a task in a few-shot learning scenario and a task's metadata, respectively.torchcross.cd
contains functions to create heads, losses and metrics for cross-domain learning experiments.
This library is still in beta. The API is potentially subject to change. Any feedback is welcome.
Installation
The library can be installed via pip:
pip install torchcross
Examples
See the examples
directory.
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
torchcross-0.0.2.tar.gz
(27.4 kB
view hashes)
Built Distribution
torchcross-0.0.2-py3-none-any.whl
(40.5 kB
view hashes)
Close
Hashes for torchcross-0.0.2-py3-none-any.whl
Algorithm | Hash digest | |
---|---|---|
SHA256 | 04d1c1b90f4bbd71c9d73f2fc1eceb9578fb71eacf1199e609577455d5e03fa3 |
|
MD5 | 39363c1f6955e9cc9f24e5325808f2b3 |
|
BLAKE2b-256 | 9d0a803e56dd1621b48cc60581cc46203d6e80281c5730a5f40a45c5afde011c |