Metadata-Version: 2.0
Name: tensorop
Version: 0.0.8.1
Summary: Simplified DL lib for simplified accelerated research
Home-page: https://github.com/prajjwal1/tensorop
Author: Prajjwal Bhargava
Author-email: prajjwalin@pm.me
License: MIT
Platform: UNKNOWN
Classifier: License :: OSI Approved :: MIT License
Classifier: Programming Language :: Python :: 3.6
Classifier: Programming Language :: Python :: 3.7
Classifier: Topic :: Scientific/Engineering
Classifier: Development Status :: 4 - Beta
Description-Content-Type: text/markdown
Requires-Dist: numpy

# tensorop
<img style="float: centre;" src="tensorop_logo.png">

[![PyPI version](https://badge.fury.io/py/tensorop.svg)](https://badge.fury.io/py/tensorop) 

Tensorop is a Deep Learning library built over Pytorch to accelerate research in the simplest manner possible without compromising on flexibility. Each module/object returned can be used with Pytorch at any moment (total sync). It can also be extended with [fastai](https://github.com/fastai/fastai).

## Getting Started

### Prerequisites

Install `pytorch` and `torchvision` from [pytorch.org](pytorch.org)
- Pytorch 1.0
- Torchvision
- Fastai (Optional: Can be extended)
- Pandas
- Numpy 

### Installing
Using with git (Recommended method)
``` 
$ git clone https://github.com/prajjwal1/tensorop
$ cd tensorop
```

Installation via Pypi
```
$ pip3 install tensorop
```
Pypi install is not being updated regularly until v0.1 comes out

### Example/Tutorials
Tutorials can be found [here](https://github.com/prajjwal1/tensorop/tree/master/nbs). They may not be updated,please refer to docs for updated API or file an issue. 
- Image Classification
    - [Example 1](https://github.com/prajjwal1/tensorop/blob/master/nbs/example1.ipynb)

## Docs
Docs can be found [here](https://github.com/prajjwal1/tensorop/tree/master/docs).

### Components (Structure)
- Model
- Dataset
- Loss
- Norm
- Torch Utils
- Trainer
- Transforms
- Ensembling 
- Utils

These are frequently changing once `v0.1` is out

## Contributing
There is so much work which needs to be done as of now, PRs are always welcome.



