How to lock layers for fine-tuning

CHDyshli / HrSegNet4CrackSegmentation

Real-time High-Resolution Neural Network with Semantic Guidance for Crack Segmentation

Apache License 2.0

49 stars 1 forks source link

How to lock layers for fine-tuning #14

Open patcharees opened 3 months ago

patcharees commented 3 months ago

Hi,

I would like to use a pretrained model and train only the last layer using a new small dataset. Is it possible to just define in the config file? How?

Thanks in advance Patcharee