https://github.com/dsergiu/pytorch-lightning-beginner

Showcase different Machine Learning Techniques using Pytorch Lightning.
https://github.com/dsergiu/pytorch-lightning-beginner

pytorch pytorch-lightning pytorch-lightning-tutorial

Last synced: 28 days ago
JSON representation

Showcase different Machine Learning Techniques using Pytorch Lightning.

• Host: GitHub
• URL: https://github.com/dsergiu/pytorch-lightning-beginner
• Owner: DSergiu
• License: mit
• Created: 2020-09-06T16:27:08.000Z (almost 6 years ago)
• Default Branch: master
• Last Pushed: 2022-11-29T21:45:30.000Z (over 3 years ago)
• Last Synced: 2025-01-29T14:16:16.404Z (over 1 year ago)
• Topics: pytorch, pytorch-lightning, pytorch-lightning-tutorial
• Language: Jupyter Notebook
• Homepage:
• Size: 2.01 MB
• Stars: 0
• Watchers: 1
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

README

          


  Pytorch Lightning Beginner






   

      

   

   

      

   

   

      

   



Showcase different ML techniques using Pytorch Lightning inspired by [pytorch-beginner](https://github.com/L1aoXingyu/pytorch-beginner).

The repository is split into `chapters` starting with basic linear regression and ending with more complex networks (e.g. RNNs, GANs). 

There is a jupyter notebook within each chapter where one can experiment on training and usage of the trained models.

## How to use?

1. Clone repo

    * `git clone https://github.com/DSergiu/pytorch-lightning-beginner`

    * `cd pytorch-lightning-beginner`

2. Create and activate virtual env

    * `python3 -m venv venv && ./venv/Scripts/activate`

3. Install dependencies

    * `pip install -r requirements.txt`

4. Start jupyter notebook

    * `jupyter notebook`

5. Open a notebook chapter and experiment (e.g. )

    * e.g. open `01-Linear-Regression/Notebook.ipynb`

## Useful training arguments

When training you can experiment with following arguments. For a full list see [Pytorch Lightning Trainer](https://pytorch-lightning.readthedocs.io/en/latest/common/trainer.html).

| Argument               | Description                                           |

|------------------------|-------------------------------------------------------|

| `--help`               | see list of arguments                                 |

| `--fast_dev_run=True`  | fast run of training                                  |

| `--max_epochs=20`      | run 20 epochs of training data set                    |

| `--accelerator=cuda`   | train on cuda GPU                                     |

| `--devices=2`          | train using 2 devices                                 |

| `--deterministic=True` | training always produces same output given same input |

## Prerequisites

You must be familiar with Python, Pytorch and Pytorch Lightning.

To understand the code please see official [Pytorch Lighning Docs](https://pytorch-lightning.readthedocs.io/en/stable/).

## Requirements

* python 3.8