Can't fine-tune Multilingual_PP-OCRv3_det_distill_train.tar

[Tanchik24]

Hello, I have a task to fine-tune the Multilingual_PP-OCRv3_det_distill_train.tar model on a French dataset with dense and small text. I would like to use fine-tuning, but the documentation suggests using this yml file - https://github.com/PaddlePaddle/PaddleOCR/blob/release/2.6/configs/det/ch_PP-OCRv3/ch_PP-OCRv3_det_cml.yml for the Multilingual_PP-OCRv3_det_distill_train.tar model. This file uses distillation, which consumes a lot of memory, so Google Colab cannot handle it. Could you please advise me on how else I can fine-tune the model trained on a large amount of data?