Sakr

can I use pretrained checkpoint to test the raw cylinder 3d we get from https://github.com/xinge008/Cylinder3D

Pseudo dice while training and validation

8

I want to ask while training in the logs I found as an example : 2024-04-23 15:22:40.711169: Epoch 195 2024-04-23 15:22:40.724555: Current learning rate: 0.00036 2024-04-23 15:23:41.803760: train_loss -0.7526 2024-04-23...

Is there any tutorial to serve the nnUNet models with torch serve

I want to increase the batch size during inference

4

I want to increase the batch size during inference with this command nnUNetv2_predict

Use the model

40

How can I use the model on the fly that gives the trained model the image and then return a variable that has the resulted image I don't want to...

file not found nnUNetTrainer__nnUNetPlans__2d/dataset.json

1

When I do predict this time it gives me this error : file not found nnUNetTrainer__nnUNetPlans__2d/dataset.json but the file is exists in the directory and I do prediction many time...

Ensemble

2

I trained 2d image for segmentation task with 2D configuration and I write that command: nnUNetv2_find_best_configuration DATASET_NAME_OR_ID -c CONFIGURATIONS but it didn't give me a command to use ensemble how...

is their other way to make the prediction with Adding batch size

1

is their other way to make the prediction with Adding batch size instead of giving the images like that: 

I want to create a python class for inference

1

I want to create a python class for inference no importance if it is long can any one help me (I want to create the usual pytorch class)