How to Fine-tune LayoutLMv3 with Annotated Documents Using PaddleOCR Part-2: Label with label-studio

Подписаться 418

Просмотров 8 тыс.

50% 1

In this tutorial, we will learn how to fine-tune LayoutLMv3 with annotated documents using PaddleOCR. LayoutLMv3 is a powerful text detection and layout analysis model that can be used to extract text from documents. PaddleOCR is an open-source OCR system that supports a variety of languages and document types.
To fine-tune LayoutLMv3 with annotated documents, we will need to:
1. PaddleOCR
2. Label-studio
3. Transformers - huggingFace
Code link : github.com/manikanthp/LayoutL...
LayoutLMv3, Fine-tune, Annotated Documents, PaddleOCR, Text Recognition, Document Layout Analysis, Computer Vision, Natural Language Processing, Deep Learning

Наука

Опубликовано:

13 июн 2023

Ссылка:

Скачать:

Готовим ссылку...

Добавить в:

Мой плейлист

Посмотреть позже

Комментарии : 18

@yeojinkim5100 Год назад

Great! I'm looking forward to your next video

@AIOdysseyhub Год назад

Thank you for the support

@WingSteels 2 месяца назад

Thank you from France you helped me a lot

@PurushothamReddy-ff6vp 2 месяца назад

hello, can you tell me how to assign the key value pairs to assign tabular data in layoutlm? or any key value pairs

@user-rx7td3lr4i 8 месяцев назад

hi, great video The issue I am facing is that I am unable to import the ocr json. Any suggestions?

@AIOdysseyhub 8 месяцев назад

Check the python src code of converting image to json file weather OCR is identifying the words/char correctly, if yes then check format in which you are creating the json, if every thing is correct reinstall label-studio, if still facing the issue please let me know Thanks for reaching out, Please subscribe to the channel for more such video Thank you 😊😊😊😊😊😊

@user-ge5wr5ue1b 10 месяцев назад

i have one doubt, like we have 100 invoices(images), then will Conv.......py file create only one json file for all that images?

@AIOdysseyhub 10 месяцев назад

yes, it will create only one json file for all the train images.

@anzias1038 6 месяцев назад

hi, good video I am facing a problem when I exported the json file from labelstudio, I got only one json file that contain details of one image.

@AIOdysseyhub 6 месяцев назад

Hi, Please check the code, In for loop you have not given the path of all images correct or looping related issue. Thank you for the support. Please subscribe the channel for more such video and support.

@rudyoactiv 9 месяцев назад

Is it possible to train it to work on 2-page documents, where the first page always has elements like "header" but the second page does not?

@AIOdysseyhub 9 месяцев назад

You train each of the page as seperately, else combined two pages in one image as train the model

@rudyoactiv 9 месяцев назад

@@AIOdysseyhub I am very new to this, does "training" always imply using several documents or can I just work with a single document, assuming all future docs that I will face will have the same layout?

@AIOdysseyhub 9 месяцев назад

@@rudyoactiv if your layout is fixed you can go with python conventional programming instead of AI models

@iSolveTechnologies-js1qj 9 месяцев назад

hi brother , submit button not working in label-studio

@iSolveTechnologies-js1qj 9 месяцев назад

label-studio submit button not working can we help out me please, in this below attached

@AIOdysseyhub 9 месяцев назад

Hi 1) Try to reinstall label-studio again 2) I have used python 3.9.0 try to check based on your python version if its having any issue. 3) Install label-studio in global python interpreter not in any env Thanks for reaching out in comments. Please subscribe the channel if you like it. will post more such videos and improve the content. Thanks again. Please let me know if you issue has solved or not. 🙂🙂