Metadata-Version: 2.1
Name: check-orientation
Version: 0.0.5
Summary: Check if image was rotated by 90, 180, 270 degrees.
Home-page: https://github.com/ternaus/check_orientation
Author: Vladimir Iglovikov
License: MIT
Platform: UNKNOWN
Classifier: License :: OSI Approved :: MIT License
Classifier: Intended Audience :: Developers
Classifier: Intended Audience :: Science/Research
Classifier: Operating System :: OS Independent
Classifier: Programming Language :: Python
Classifier: Programming Language :: Python :: 3
Classifier: Topic :: Software Development :: Libraries
Classifier: Topic :: Software Development :: Libraries :: Python Modules
Description-Content-Type: text/markdown
Requires-Dist: albumentations
Requires-Dist: iglovikov-helper-functions
Requires-Dist: timm
Requires-Dist: torch
Provides-Extra: test
Requires-Dist: pytest ; extra == 'test'

# Check orientation

![](https://habrastorage.org/webt/hb/ru/ii/hbruiiuortx05lxfuokzw1skheu.jpeg)

Models to check if image was rotated by 0, 90, 180, 270 degrees.

## Installation
`pip install -U check_orientation`

### Example inference

Colab notebook with the example: [![Open In Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/drive/1HIGM_b8rH20N414ROZB0HN9w5p4ko2Bd?usp=sharing)

## Training

### Define the config.
Example at [check_orientation/configs](check_orientation/configs)

### Define the environmental variable `TRAIN_IMAGE_PATH` that points to the folder with train dataset.
Example:
```bash
export TRAIN_IMAGE_PATH=<path to the tranining folder>
```

### Define the environmental variable `VAL_IMAGE_PATH` that points to the folder with validation dataset.
Example:
```bash
export VAL_IMAGE_PATH=<path to the validation folder>
```

### Training
```
python -m check_orientation.train -c <path to config>
```

### Inference

```bash
python -m torch.distributed.launch --nproc_per_node=<num_gpu> check_orientation/inference.py \
                                   -i <path to images> \
                                   -c <path to config> \
                                   -w <path to weights> \
                                   -o <output-path> \
                                   --fp16
```

### Pre-trained models
Models were pre-trained on the [OpenImages dataset](https://storage.googleapis.com/openimages/web/index.html).

| Models        | Validation accuracy | Config file  | Weights |
| ------------- |:--------------------:| :------------:| :------: |
| swsl_resnext50_32x4d|0.9128| [Link](check_orientation/configs/2020-11-16.yaml)| [Link](https://github.com/ternaus/check_orientation/releases/download/v0.0.3/2020-11-16_resnext50_32x4d.zip)|


