search

arvindrajan92 / DTrOCR

A PyTorch implementation of DTrOCR: Decoder-only Transformer for Optical Character Recognition
MIT License
79 stars 8 forks source link

Model Finetuning #14

Closed donjuanpond closed 3 months ago

donjuanpond commented 3 months ago

Just wanted to ask, are you working on creating a finetuning script for this model? If so, when do you plan on releasing it?

arvindrajan92 commented 3 months ago

hello @donjuanpond, there is no script as such, but there is a finetuning example here - hope you'll find this useful. also note that the model has not been pre-trained, hence you may not get the expected performance.

currently, i am doing the groundwork to pre-train on printed and handwritten texts. script for fine-tuning would come later.

arvindrajan92 commented 3 months ago

@donjuanpond are you happy for me to close this issue? or do you need assistance?

donjuanpond commented 3 months ago

Sorry for not closing myself, everything seems good!