InnerEye-DeepLearning
InnerEye-DeepLearning copied to clipboard
microsoft
Medical Imaging Deep Learning library to train and deploy 3D segmentation models on Azure Machine Learning
[AB#3932](https://innereye.visualstudio.com/60ce1777-00d6-4015-82bc-488a0c00202f/_workitems/edit/3932)
Describing how to use TRE to run InnerEye bits [AB#5940](https://innereye.visualstudio.com/60ce1777-00d6-4015-82bc-488a0c00202f/_workitems/edit/5940)
recovery_start_epoch is not settable. We use it to determine the list of training epochs - but even that could be dead code. Verify and clean up [AB#4535](https://innereye.visualstudio.com/60ce1777-00d6-4015-82bc-488a0c00202f/_workitems/edit/4535)
To check in particular regressions when packages get updated. [AB#4223](https://innereye.visualstudio.com/60ce1777-00d6-4015-82bc-488a0c00202f/_workitems/edit/4223)
It would be good to have the HelloWorld model in a form that it can be used also as an example for deployment. At present, that is not possible, because...
Training a model with this params: --model=HeadAndNeck --train=True --cluster=training-nd40-v2 --run_recovery_id=jaalvare_bigcluster:jaalvare_bigcluster_1607291058_70ac48b7 --start_epoch=120 Results on: 2021-01-05T13:13:20Z ERROR Model training/testing failed. Exception: CUDA out of memory. Tried to allocate 12.74 GiB (GPU 0;...
From the present thumbnails, it is not clear which pixels will never be included in training. Can we highlight those in a different color? [AB#3939](https://innereye.visualstudio.com/60ce1777-00d6-4015-82bc-488a0c00202f/_workitems/edit/3939)
Hi guys, This was a suggestion that I brought up to the InnerEye-CreateDataSet repo, but I was advised to repost it here: _Suggestion:_ For the case where the image+segmentation sets...
Run IDs are now unique within the workspace, and the API no longer needs an experiment in order to retrieve a run. This will simplify the code. [AB#4536](https://innereye.visualstudio.com/60ce1777-00d6-4015-82bc-488a0c00202f/_workitems/edit/4536)
#554 disabled `test_recovery_on_2_nodes` because the run launched by the test would hang and would not be marked complete (even though the InnerEye script itself finished running). We need to triage...
