Ecosyste.ms: Awesome
An open API service indexing awesome lists of open source software.
https://github.com/idealo/imageatm
Image classification for everyone.
https://github.com/idealo/imageatm
computer-vision deep-learning image-classification keras machine-learning neural-network tensorflow transfer-learning
Last synced: 6 days ago
JSON representation
Image classification for everyone.
- Host: GitHub
- URL: https://github.com/idealo/imageatm
- Owner: idealo
- License: apache-2.0
- Created: 2019-03-18T10:32:00.000Z (over 5 years ago)
- Default Branch: master
- Last Pushed: 2023-12-19T09:36:57.000Z (11 months ago)
- Last Synced: 2024-07-15T15:37:37.606Z (4 months ago)
- Topics: computer-vision, deep-learning, image-classification, keras, machine-learning, neural-network, tensorflow, transfer-learning
- Language: Python
- Homepage: https://idealo.github.io/imageatm/
- Size: 13.2 MB
- Stars: 214
- Watchers: 22
- Forks: 49
- Open Issues: 9
-
Metadata Files:
- Readme: README.md
- Contributing: CONTRIBUTING.md
- License: LICENSE
Awesome Lists containing this project
README
# Image ATM (Automated Tagging Machine)
[](https://travis-ci.org/idealo/imageatm)
[](https://github.com/idealo/imageatm/blob/master/LICENSE)
Image ATM is a one-click tool that automates the workflow of a typical image classification pipeline in an opinionated way, this includes:
- Preprocessing and validating input images and labels
- Starting/terminating cloud instance with GPU support
- Training
- Model evaluation
Read the documentation at: [https://idealo.github.io/imageatm/](https://idealo.github.io/imageatm/)
Image ATM is compatible with Python 3.6 and is distributed under the Apache 2.0 license.
## Installation
There are two ways to install Image ATM:
* Install Image ATM from PyPI (recommended):
```
pip install imageatm
```
* Install Image ATM from the GitHub source:
```
git clone https://github.com/idealo/imageatm.git
cd imageatm
python setup.py install
```
## Usage
#### Train with CLI
Run this in your terminal
```
imageatm pipeline config/config_file.yml
```
#### Train without CLI
Run the data preparation:
``` python
from imageatm.components import DataPrep
dp = DataPrep(
samples_file = 'sample_configfile.json',
image_dir = 'sample_dataset/',
job_dir='sample_jobdir/'
)
dp.run(resize=True)
```
Run the training:
``` python
from imageatm.components import Training
trainer = Training(image_dir=dp.image_dir, job_dir=dp.job_dir)
trainer.run()
```
Run the evaluation:
``` python
from imageatm.components import Evaluation
evaluator = Evaluation(image_dir=dp.image_dir, job_dir=dp.job_dir)
evaluator.run()
```
## Test
Test execution is triggered by these commands:
```
pip install -e ".[tests, docs]"
pytest -vs --cov=imageatm --show-capture=no --disable-pytest-warnings tests/
```
## Transfer learning
The following pretrained CNNs from Keras can be used for transfer learning in Image-ATM:
- Xception
- VGG16
- VGG19
- ResNet50, ResNet101, ResNet152
- ResNet50V2, ResNet101V2, ResNet152V2
- ResNeXt50, ResNeXt101
- InceptionV3
- InceptionResNetV2
- MobileNet
- MobileNetV2
- DenseNet121, DenseNet169, DenseNet201
- NASNetLarge, NASNetMobile
Training is split into two phases, at first only the last dense layer gets
trained, and then all layers are trained.
For each phase the **learning rate is reduced** after a patience period if no
improvement in validation accuracy has been observed. The patience period
depends on the average number of samples per class (*n_per_class*):
- if *n_per_class* < 200: patience = 5 epochs
- if *n_per_class* >= 200 and < 500: patience = 4 epochs
- if *n_per_class* >= 500: patience = 2 epochs
**Training is stopped early** after a patience period that is three times
the learning rate patience to allow for two learning rate adjustments
before stopping training.
## Contribute
We welcome all kinds of contributions. See the [Contribution](CONTRIBUTING.md) guide for more details.
#### Bump version
To bump up the version, use
```
bumpversion {part} setup.py
```
## Cite this work
Please cite Image ATM in your publications if this is useful for your research. Here is an example BibTeX entry:
```
@misc{idealods2019imageatm,
title={Image ATM},
author={Christopher Lennan and Malgorzata Adamczyk and Gunar Maiwald and Dat Tran},
year={2019},
howpublished={\url{https://github.com/idealo/imageatm}},
}
```
## Maintainers
* Christopher Lennan, github: [clennan](https://github.com/clennan)
* Malgorzata Adamczyk, github: [gosia-malgosia](https://github.com/gosia-malgosia)
* Gunar Maiwald: github: [gunarmaiwald](https://github.com/gunarmaiwald)
* Dat Tran, github: [datitran](https://github.com/datitran)
## Copyright
See [LICENSE](LICENSE) for details.
## TO-DOs:
- We are currently using Keras 2.2. The plan is to use tf.keras once TF 2.0 is out. Currently tf.keras is buggy,
especially with model saving/loading (https://github.com/tensorflow/tensorflow/issues/22697)