How to finetune using pretrained model?

caiyuanhao1998 / Retinexformer

"Retinexformer: One-stage Retinex-based Transformer for Low-light Image Enhancement" (ICCV 2023) & (NTIRE 2024 Challenge)

https://arxiv.org/abs/2303.06705

MIT License

828 stars 64 forks source link

How to finetune using pretrained model? #94

Closed vishxlyadav closed 2 months ago

vishxlyadav commented 2 months ago

Thanks for the great work. I want to finetune the model on my dataset using the pretrained weights on Adobe5K dataset. I found that the .yml file for options has an argument of path --> pretrain_network_g. Is it enough to just give the path of the pretrained model here for finetuning. Also please mention the steps to finetune the model in the readme. Thanks!!

caiyuanhao1998 commented 2 months ago

Yes, that's it. And remember to modify the learning rate and lr scheduler.

If you find our repo useful, please help us star it. Thanks :)