search

clovaai / donut

Official Implementation of OCR-free Document Understanding Transformer (Donut) and Synthetic Document Generator (SynthDoG), ECCV 2022
https://arxiv.org/abs/2111.15664
MIT License
5.53k stars 444 forks source link

Are the parameters of swinTransformer trained during fine-tuning? #208

Open DYF-AI opened 1 year ago

DYF-AI commented 1 year ago

Are the parameters of swinTransformer trained during fine-tuning?

gwkrsrch commented 1 year ago

@DYF-AI Yes, during the fine-tuning phase, the vision encoder is trained as well.

DYF-AI commented 1 year ago

@DYF-AI Yes, during the fine-tuning phase, the vision encoder is trained as well.

Thank you for your answer, and one more question, is there any code that provides training the Vision Encoder?