Metadata-Version: 2.1
Name: pytorchlight
Version: 0.0.1
Summary: Easy-to-run Application based on PyTorch
Home-page: https://github.com/mechanicalsea/pytorch-light
Author: Rui Wang
Author-email: wangrui_key@163.com
License: BSD
Platform: UNKNOWN
Classifier: Programming Language :: Python :: 3.5
Classifier: License :: OSI Approved :: BSD License
Classifier: Operating System :: Microsoft :: Windows :: Windows 10
Description-Content-Type: text/markdown
Requires-Dist: logging (==0.5.1.2)
Requires-Dist: numpy (==1.15.2)
Requires-Dist: matplotlib (==2.1.1)
Requires-Dist: torch (==1.0.1)
Requires-Dist: torchvision (==0.2.2)

# Pytorch Light
Pytorch Light is a Python package that provides several tools about modelling and applications.

There are some features as follows:

- [Model](#model)
- [Data](#data)
- [Log](#log)
- [Visualization](#visualization)

[requirements.txt](requirements.txt) is also provided.

Next feature is going to be achieved as soon as possible.

- Adversarial Learning

And some applications are coming:

- Computer Vision
- Natural Language
- Rereinforcement Learning

**Note** that these features are only validated on cpu.

## Model

Application for model.

Features:

1. train
2. evaluate
3. batch loss and batch metric
4. model wrapper

## Data

Application for data.

Features:

1. Numpy.ndarray to/from Torch.tensor
2. Numpy.ndarray to/from Torch.Dataset
3. Numpy.ndarray to/from Torch.DataLoader

## Log

Application for log.

Features:

1. debug
2. info
3. warn
4. error
5. critical

## Visualization

Application for visualization based on numpy.ndarray.

Require:

- In the notebook mode, `%matplotlib inline` should be implement first.

Features:

1. learning curves
2. batch curves
3. one or multi images


