https://github.com/andreaschandra/ocr-handwritten
An end-to-end model development for handwritten OCR
https://github.com/andreaschandra/ocr-handwritten
Last synced: about 1 year ago
JSON representation
An end-to-end model development for handwritten OCR
- Host: GitHub
- URL: https://github.com/andreaschandra/ocr-handwritten
- Owner: andreaschandra
- License: mit
- Created: 2023-02-06T02:03:01.000Z (over 3 years ago)
- Default Branch: main
- Last Pushed: 2023-02-09T00:42:09.000Z (over 3 years ago)
- Last Synced: 2025-02-12T09:49:05.741Z (over 1 year ago)
- Language: Jupyter Notebook
- Size: 11.1 MB
- Stars: 0
- Watchers: 2
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
# ocr-handwritten
An end-to-end model development for handwritten OCR
Experiment Tracking [here](https://app.clear.ml/projects/d60f182ac5104605afd7f1c45ff2a927/experiments/f2fb3eb021f8408b9da26fd91ac65c20/output/execution)
# Requirements
Please make sure that you have docker installed on you device.
# Weights
Download craft model from [here](https://drive.google.com/file/d/1Jk4eGD7crsqCCg9C9VjCLkMN3ze8kutZ/view) and place it to `scripts/model/`
Download trocr model from [here](https://drive.google.com/drive/folders/1TR_Hrre_SqXKICyWTUeG799NG-OZVnw7?usp=sharing) and extract the model to `scripts/model/` and rename as `trocr-finetuned-augmented`
## Run the program
First, build docker image
`docker-compose build`
the, run the container
`docker-compose up`