https://github.com/ternaus/iglovikov_segmentation
Semantic segmentation pipeline using Catalyst.
https://github.com/ternaus/iglovikov_segmentation
cityscapes computer-vision deep-learning python semantic-segmentation
Last synced: 9 months ago
JSON representation
Semantic segmentation pipeline using Catalyst.
- Host: GitHub
- URL: https://github.com/ternaus/iglovikov_segmentation
- Owner: ternaus
- License: mit
- Created: 2019-10-10T04:35:44.000Z (over 6 years ago)
- Default Branch: master
- Last Pushed: 2020-04-03T02:16:39.000Z (over 6 years ago)
- Last Synced: 2025-04-11T16:25:12.546Z (about 1 year ago)
- Topics: cityscapes, computer-vision, deep-learning, python, semantic-segmentation
- Language: Python
- Homepage:
- Size: 86.9 KB
- Stars: 20
- Watchers: 3
- Forks: 1
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- Contributing: docs/contributing.rst
- License: LICENSE
Awesome Lists containing this project
README
# Installation
Run `pip install -r requirements.txt` to install all requred libraries.
# Data preparation
## Cityscapes
1. Use [this](https://github.com/ternaus/iglovikov_helper_functions/tree/master/iglovikov_helper_functions/data_processing/cityscapes)
guide to prepare the data.
2. Put prepared data to the `data` folder or modify path to the train and test files in the config file.
# Training
Almost all trainining parameters are specified in the config file. You may check examples in the configs folder.
You may star training, running:
`python -m src.train `
# Inference
Run:
```
python -m src.inference -c \
-i \
-o
-w \
-b \
-j
```
# Contributing
To contribute follow [this guide.](docs/contributing.rst)