https://github.com/jyotidabass/document_text_recognition

https://github.com/jyotidabass/document_text_recognition

Last synced: 3 months ago
JSON representation

• Host: GitHub
• URL: https://github.com/jyotidabass/document_text_recognition
• Owner: jyotidabass
• License: apache-2.0
• Created: 2022-03-09T11:02:36.000Z (over 2 years ago)
• Default Branch: main
• Last Pushed: 2022-03-09T11:12:43.000Z (over 2 years ago)
• Last Synced: 2024-06-28T11:32:55.251Z (5 months ago)
• Language: Python
• Size: 195 KB
• Stars: 2
• Watchers: 1
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md
  • Changelog: changelog.rst
  • Contributing: CONTRIBUTING.md
  • License: LICENSE
  • Code of conduct: CODE_OF_CONDUCT.md

Awesome Lists containing this project

README

        
# Text recognition

The sample training script was made to train text recognition model with docTR.

## Setup

First, you need to install `doctr` (with pip, for instance)

```shell

pip install -e . --upgrade

pip install -r references/requirements.txt

```

## Usage

You can start your training in TensorFlow:

```shell

python references/recognition/train_tensorflow.py crnn_vgg16_bn --train_path path/to/your/train_set --val_path path/to/your/val_set  --epochs 5

```

or PyTorch:

```shell

python references/recognition/train_pytorch.py crnn_vgg16_bn --train_path path/to/your/train_set --val_path path/to/your/val_set --epochs 5 --device 0

```

## Data format

You need to provide both `train_path` and `val_path` arguments to start training.

Each of these paths must lead to a 2-elements folder:

```shell

├── images

    ├── img_1.jpg

    ├── img_2.jpg

    ├── img_3.jpg

    └── ...

├── labels.json

```

The JSON files must contain word-labels for each picture as a string.

The order of entries in the json does not matter.

```shell

labels = {

    'img_1.jpg': 'I',

    'img_2.jpg': 'am',

    'img_3.jpg': 'a',

    'img_4.jpg': 'Jedi',

    'img_5.jpg': '!',

    ...

}

```

## Advanced options

Feel free to inspect the multiple script option to customize your training to your own needs!

```python

python references/recognition/train_pytorch.py --help

```