![](https://img-blog.csdnimg.cn/20201014180756919.png?x-oss-process=image/resize,m_fixed,h_64,w_64)
nnUnet
文章平均质量分 87
Cierlly
继续搞啊!
展开
-
expected_epoch_times
IntroductionTrainings can take some time. A well-running training setup is essential to get the most of nnU-Net. nnU-Net does notrequire any fancy hardware, just a well-balanced system. We recommend at least 32 GB of RAM, 6 CPU cores (12 threads),SSD st转载 2021-03-27 13:03:54 · 259 阅读 · 0 评论 -
Extending/Changing nnU-Net
Extending/Changing nnU-NetTo use nnU-Net as a framework and make changes to its components, please make sure to install it with the git cloneand pip install -e . commands so that a local copy of the code is created.Changing components of nnU-Net needs t转载 2021-03-27 13:03:09 · 296 阅读 · 0 评论 -
edit_plans_files
Changing the plans files grants you a lot of flexibility: You can depart from nnU-Net’s default configuration and playwith different U-Net topologies, batch sizes and patch sizes. It is a powerful tool!To better understand the components describing the n转载 2021-03-27 13:01:40 · 140 阅读 · 0 评论 -
custom_spacing
Sometimes you want to set custom target spacings. This is done by creating a custom ExperimentPlanner.Let’s run this with the Task002_Heart example from the Medical Segmentation Decathlon. This dataset is not too largeand working with it is therefore a b转载 2021-03-27 13:00:59 · 110 阅读 · 0 评论 -
custom_preprocessing
When you would like to interfere with the way resampling during preprocessing is handled or you would like to implementa custom normalization scheme, you need to create a new custom preprocessor class and an ExperimentPlanner to go alongwith it. While th转载 2021-03-27 13:00:21 · 127 阅读 · 0 评论 -
Common Issues and their Solutions
Common Issues and their SolutionsRuntimeError: Expected scalar type half but found floatThis can happen when running inference (or training) with mixed precision enabled on older GPU hardware. It pointsto some operation not being implemented in half pre转载 2021-03-27 12:59:00 · 471 阅读 · 0 评论 -
Data format for Inference
Data format for InferenceThe data format for inference must match the one used for the raw data (specifically, the images must be in exactlythe same format as in the imagesTr folder). As before, the filenames must start with aunique identifier, followed转载 2021-03-27 12:58:16 · 129 阅读 · 0 评论 -
Setting up Paths
Setting up PathsnnU-Net relies on environment variables to know where raw data, preprocessed data and trained model weights are stored.To use the full functionality of nnU-Net, the following three environment variables must be set:nnUNet_raw_data_base转载 2021-03-27 12:57:02 · 163 阅读 · 0 评论 -
Dataset conversion instructions
Dataset conversion instructionsnnU-Net requires the raw data to be brought into a specific format so that it know how to read and interpret it. Thisformat closely, but not entirely, follows the format used by theMedical Segmentation Decathlon (MSD).The转载 2021-03-25 22:29:08 · 226 阅读 · 0 评论 -
nnUNet
[2020_10_21] Update: We now have documentation for common questions andcommon issues. We now also provide reference epoch times forseveral datasets and tips on how to identify bottlenecks.Please read these documents before opening a new issue!nnU-NetI转载 2021-03-25 22:27:47 · 1377 阅读 · 1 评论