Hi!
It seems that the command you have shared to finetune the model gives an error because --resume does not accept an argument of model path. --resume is just a boolean flag which will resume training from the output directory if a checkpoint is found there. I think MODEL.WEIGHTS is the config parameter we'll have to change to load a specific checkpoint. Also, I think we'll have to pass in --config in the command to specify which YAML file to load.
Hi!
It seems that the command you have shared to finetune the model gives an error because
--resumedoes not accept an argument of model path.--resumeis just a boolean flag which will resume training from the output directory if a checkpoint is found there. I thinkMODEL.WEIGHTSis the config parameter we'll have to change to load a specific checkpoint. Also, I think we'll have to pass in--configin the command to specify which YAML file to load.