yuji96 / train-tesseract Goto Github PK

View Code? Open in Web Editor NEW

This project forked from guiem/train-tesseract

0.0 0.0 0.0 2.65 MB

Dockerized example to train Tesseract v. 4

Shell 4.31% Python 2.64% Dockerfile 10.98% Makefile 82.08%

train-tesseract's People

Contributors

train-tesseract's Issues

Do not clone tesstrain but download it as zip

train-tesseract/start.sh

Lines 10 to 11 in 457452b

    
           # TODO: change to zip 
        
           git clone --depth=1 https://github.com/tesseract-ocr/tesstrain.git

I don't want .git in tesstrain

hard cording. DATA_DIR should be read from args.txt

train-tesseract/start.sh

Lines 26 to 27 in 457452b

    
           # TODO: use DATA_DIR 
        
           if [ ! -d data/langdata ]; then

train-tesseract/args.txt

Line 7 in 457452b

    
           # DATA_DIR           # Data directory for output files, proto model, start model, etc. Default: data

replace ppa:deadsnakes/ppa with universe

I have same concern.
https://askubuntu.com/a/1398569

train-tesseract/Dockerfile

Lines 40 to 41 in 457452b

    
           RUN apt install -y software-properties-common && add-apt-repository -y ppa:deadsnakes/ppa && \ 
        
               apt update && apt install -y python3.11

universe also have python3.11
http://archive.ubuntu.com/ubuntu/pool/universe/p/python3.11/

Is it easier to install make-4.4 if using Debian?

apt install -y make=4.4
or
apt install -y ???-backports make

MODEL_NAME should not be defined in start.sh but read from args.txt

train-tesseract/start.sh

Line 6 in 457452b

MODEL_NAME=sample

train-tesseract/args.txt

Line 1 in 457452b

MODEL_NAME=sample # Name of the model to be built. Default: foo

Why eng_best.traineddata is not found?

START_MODEL is eng_best, but tesseract search eng.traineddata during make training.

train-tesseract/args.txt

Line 2 in 457452b

START_MODEL=eng_best # Name of the model to continue from. Default: ''

here is temporary repairs.

train-tesseract/start.sh

Line 24 in 457452b

    
           ln -s /usr/local/share/tessdata/eng_best.traineddata /usr/local/share/tessdata/eng.traineddata

Maybe here?, Can I use --lang $(START_MODEL)?
https://github.com/tesseract-ocr/tesstrain/blob/main/Makefile#L244-L263

yuji96 / train-tesseract Goto Github PK

train-tesseract's People

Contributors

train-tesseract's Issues

Do not clone tesstrain but download it as zip

hard cording. DATA_DIR should be read from args.txt

replace ppa:deadsnakes/ppa with universe

Is it easier to install make-4.4 if using Debian?

MODEL_NAME should not be defined in start.sh but read from args.txt

Why eng_best.traineddata is not found?

Recommend Projects

React

Vue.js

Typescript

TensorFlow

Django

Laravel

D3

Recommend Topics

javascript

web

server

Machine learning

Visualization

Game

Recommend Org

Facebook

Microsoft

Google

Alibaba

D3

Tencent

	# TODO: change to zip
	git clone --depth=1 https://github.com/tesseract-ocr/tesstrain.git

	RUN apt install -y software-properties-common && add-apt-repository -y ppa:deadsnakes/ppa && \
	apt update && apt install -y python3.11