Remove LoRA layers from training SAMed

[TBOO-Y]

Hello, I am training and fine-tuning SAMed on the Cityscapes dataset (related to this issue: #12 ) and I wish to completely remove the LoRA layers during the training process as I don't want to use LoRA fine-tuning on the model. Would it be possible to give me directions on this? Looking forward to your assistance! Thank you!

[theneao]

训练为什么要删除lora，那不就是正常SAM了吗