Can you share the finetuning code
Hey DILHTWD, this is an amazing project and EasyOCR is miles ahead of Pytesseract in terms of speed. Can you please also share the code that you finetuned the model on ?
Dear Rushikesh, sure we can provide the fine-tuning code. However, I am a bit confused about your comment regarding EasyOCR. We trained a YOLO model on the DocLayNet data set instead. Best!
Yes, the other implementations of finetuning/ inference of doclaynet are almost always using pytessearct. Yours is the only one I could find that uses easyocr which is at least for this project is infinitely better. I want to finetune yolov8 on my dataset so I thought I could use the code you used for finetuning on doclaynet.
Thanks :)
Hello
@DILHTWD
,
First of all, thanks for sharing this model, quite useful.
I am working on training a model in my own custom dataset for layout segmentation and UI understanding and, therefore, I would be really glad if you can share the code (or simply the details) for fine-tunning.
Thanks! All best!