Question about how to fine-tune the table-structure-recognition model.

[YingxuanW]

My need is to fine-tune this table-structure-recognition model, but at the moment I only have some images and don't know what to do with them. Which tools should I use to process these images in a dataset format suitable for fine-tuning?

It would be nice to have a detailed process.

Thanks!

[mazeem-arbisoft]

• You will need to label the data with classes listed here
• The data format needs to be in Pascal VOC, methods to output annotations in this format are implemented here
• You can try to use available online tools (e.g. Labelimg)
• We've wrote a tool that outputs data ready for fine-tuning this model, but that is not open source yet.

[YingxuanW]

Azeem Iftikhar (@mazeem-arbisoft) Thanks a lot! Looking forward to your open source tool!