difference between model_best-v1.ckpt, model_best-v2.ckpt and model_best.ckpt

MIC-DKFZ / nnDetection

nnDetection is a self-configuring framework for 3D (volumetric) medical object detection which can be applied to new data sets without manual intervention. It includes guides for 12 data sets that were used to develop and evaluate the performance of the proposed method.

Apache License 2.0

538 stars 92 forks source link

difference between model_best-v1.ckpt, model_best-v2.ckpt and model_best.ckpt #269

Open Rajesh-ParaxialTech opened 3 weeks ago

Rajesh-ParaxialTech commented 3 weeks ago

Hello

After training an nnDetection model for several epochs I could see (see figure below) the following trained models in the "det_models" folder. I could get what "model_best.ckpt" and "model_last.ckpt" are. But I could not make out what model_best-v1.ckpt and model_best-v2.ckpt are.

Could anyone please explain ?

Rajesh

mibaumgartner commented 3 weeks ago

Dear @Rajesh-ParaxialTech ,

v1 and v2 are models which are usually created when continuing, restarting the training etc. They are generally the same as the normal _best mode.

Best, Michael

Rajesh-ParaxialTech commented 3 weeks ago

Thanks a lot Michael