https://github.com/durandtibo/segmentation_sol
https://github.com/durandtibo/segmentation_sol
Last synced: about 2 months ago
JSON representation
- Host: GitHub
- URL: https://github.com/durandtibo/segmentation_sol
- Owner: durandtibo
- License: mit
- Created: 2017-11-30T08:41:55.000Z (over 7 years ago)
- Default Branch: master
- Last Pushed: 2018-01-12T21:06:32.000Z (over 7 years ago)
- Last Synced: 2025-02-16T04:21:51.077Z (4 months ago)
- Language: Python
- Size: 5.82 MB
- Stars: 1
- Watchers: 3
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
# segmentation_sol
## Installation
- Install `conda` (https://conda.io/miniconda.html) or update it
```
conda update conda
```
- Tp create a new environment `segsol`, run the command:
```
conda create -n segsol python=3.6
```
- To activate this environment, run the command:
```
source activate segsol
```
- To install `pytorch` package (http://pytorch.org/), run the command:
```
conda install pytorch torchvision -c pytorch
```
- To install other package, run the command:
```
pip install tqdm
```
- To clone this repository, run the command:
```
git clone https://github.com/durandtibo/segmentation_sol.git
```
## Evaluation
From the folder `segmentation_sol`, to predict the mask of one image, run the command:
```
python -m segmentation.main --image data/test.JPG --max_size 1000 --output_dir outputs
```
From the folder `segmentation_sol`, to predict the mask of all the images in a directory, run the command:
```
python -m segmentation.main --image data --max_size 1000 --output_dir outputs
```
List of options
- `image`: the filename or the directory of the image
- `checkpoint_dir`: the directory where the checkpoint is
- `output_dir`: the directory where to save the results
- `max_size`: the size of the smaller edge of the image