Adding an example for fine-tuning both detector & recognizer using an your own dataset

[abdulmalik29]

I have been working on a projects that required to fine-tune both models to our numbers dataset. I have made some helper functions to convert the dataset from YOLO Keras format to the one used in training the models. I would like to provide this example to.

[DancheG14]

Это было бы крайне полезно

[asif-ca]

@abdulmalik29 Yes, It would be extremely helpful please add.

Thanks

[Terrance-Whitehurst]

I am also searching for clear documentation on how to train the keras-ocr recognizer on customer image dataset.

[seansteel3]

@Terrance-Whitehurst and @abdulmalik29 is this source basically what you're looking for?
https://keras-ocr.readthedocs.io/en/latest/examples/end_to_end_training.html

If so then all you need to do is get your data in the same format as the guide and you're ready to retrain. I also would take a look at these as well:
#231
#218