ccHuang

I changed to `offline_evaluate()` by using ``` model_path='train_log/unet3d/model-5' pred = OfflinePredictor(PredictConfig( model=get_model(modelType="inference"), session_init=get_model_loader(model_path), input_names=['image'], output_names=get_model_output_names())) ``` something goes wrong, the log shows: [1119 19:40:30 @sessinit.py:117] Restoring checkpoint from train_log/unet3d/model-5 ......

I see. Thank you. Finally, I found a solution to avoid OOM: ``` import threading thread = threading.Thread(target=self._eval(), name='self._eval') thread.start() thread.join() # wait threading to finish to close it to...

Hi @tkuanlun350 , now I'm adapting your code to LiTS (for liver segmentation) challenge. The 3D CT volume is much larger than BRATS, i.e. 512x512xn (n = 100~1000). In this...

@mini-Shark Yes. I found the reason resides in `QueueInput(get_train_dataflow())`. `get_train_dataflow()`, the process from 'loading data from hard disk' to 'get data preprocessed for queue', is on going all the time....

I have no idea using tensorpack.

@tkuanlun350 It's a different problem. I think @mini-Shark should also set 'NO_CACHE = True'. The problem is that if the online evaluation takes long time(e.g. we have half of the...

@tkuanlun350 Thank you.

@pieper Thank you for sharing the codes. But some errors raised when running the codes according to the log. And the obtained .nrrd files, input into pyradiomics's featureextractor, resulted in...

@JoostJM It works well now. Thank you so much.

DICOM RTSTRUCT can be saved as .nrrd file in Slicer software: my version: the most recent nightly 4.7.0-2017-09-10 r26368. prerequisites: install extension 'slicerRT'. Also recommend to install extension 'slicerRadiomics' which...