feat: init

nnUNet/LICENSE

@@ -0,0 +1,201 @@
+                                 Apache License
+                           Version 2.0, January 2004
+                        http://www.apache.org/licenses/
+
+   TERMS AND CONDITIONS FOR USE, REPRODUCTION, AND DISTRIBUTION
+
+   1. Definitions.
+
+      "License" shall mean the terms and conditions for use, reproduction,
+      and distribution as defined by Sections 1 through 9 of this document.
+
+      "Licensor" shall mean the copyright owner or entity authorized by
+      the copyright owner that is granting the License.
+
+      "Legal Entity" shall mean the union of the acting entity and all
+      other entities that control, are controlled by, or are under common
+      control with that entity. For the purposes of this definition,
+      "control" means (i) the power, direct or indirect, to cause the
+      direction or management of such entity, whether by contract or
+      otherwise, or (ii) ownership of fifty percent (50%) or more of the
+      outstanding shares, or (iii) beneficial ownership of such entity.
+
+      "You" (or "Your") shall mean an individual or Legal Entity
+      exercising permissions granted by this License.
+
+      "Source" form shall mean the preferred form for making modifications,
+      including but not limited to software source code, documentation
+      source, and configuration files.
+
+      "Object" form shall mean any form resulting from mechanical
+      transformation or translation of a Source form, including but
+      not limited to compiled object code, generated documentation,
+      and conversions to other media types.
+
+      "Work" shall mean the work of authorship, whether in Source or
+      Object form, made available under the License, as indicated by a
+      copyright notice that is included in or attached to the work
+      (an example is provided in the Appendix below).
+
+      "Derivative Works" shall mean any work, whether in Source or Object
+      form, that is based on (or derived from) the Work and for which the
+      editorial revisions, annotations, elaborations, or other modifications
+      represent, as a whole, an original work of authorship. For the purposes
+      of this License, Derivative Works shall not include works that remain
+      separable from, or merely link (or bind by name) to the interfaces of,
+      the Work and Derivative Works thereof.
+
+      "Contribution" shall mean any work of authorship, including
+      the original version of the Work and any modifications or additions
+      to that Work or Derivative Works thereof, that is intentionally
+      submitted to Licensor for inclusion in the Work by the copyright owner
+      or by an individual or Legal Entity authorized to submit on behalf of
+      the copyright owner. For the purposes of this definition, "submitted"
+      means any form of electronic, verbal, or written communication sent
+      to the Licensor or its representatives, including but not limited to
+      communication on electronic mailing lists, source code control systems,
+      and issue tracking systems that are managed by, or on behalf of, the
+      Licensor for the purpose of discussing and improving the Work, but
+      excluding communication that is conspicuously marked or otherwise
+      designated in writing by the copyright owner as "Not a Contribution."
+
+      "Contributor" shall mean Licensor and any individual or Legal Entity
+      on behalf of whom a Contribution has been received by Licensor and
+      subsequently incorporated within the Work.
+
+   2. Grant of Copyright License. Subject to the terms and conditions of
+      this License, each Contributor hereby grants to You a perpetual,
+      worldwide, non-exclusive, no-charge, royalty-free, irrevocable
+      copyright license to reproduce, prepare Derivative Works of,
+      publicly display, publicly perform, sublicense, and distribute the
+      Work and such Derivative Works in Source or Object form.
+
+   3. Grant of Patent License. Subject to the terms and conditions of
+      this License, each Contributor hereby grants to You a perpetual,
+      worldwide, non-exclusive, no-charge, royalty-free, irrevocable
+      (except as stated in this section) patent license to make, have made,
+      use, offer to sell, sell, import, and otherwise transfer the Work,
+      where such license applies only to those patent claims licensable
+      by such Contributor that are necessarily infringed by their
+      Contribution(s) alone or by combination of their Contribution(s)
+      with the Work to which such Contribution(s) was submitted. If You
+      institute patent litigation against any entity (including a
+      cross-claim or counterclaim in a lawsuit) alleging that the Work
+      or a Contribution incorporated within the Work constitutes direct
+      or contributory patent infringement, then any patent licenses
+      granted to You under this License for that Work shall terminate
+      as of the date such litigation is filed.
+
+   4. Redistribution. You may reproduce and distribute copies of the
+      Work or Derivative Works thereof in any medium, with or without
+      modifications, and in Source or Object form, provided that You
+      meet the following conditions:
+
+      (a) You must give any other recipients of the Work or
+          Derivative Works a copy of this License; and
+
+      (b) You must cause any modified files to carry prominent notices
+          stating that You changed the files; and
+
+      (c) You must retain, in the Source form of any Derivative Works
+          that You distribute, all copyright, patent, trademark, and
+          attribution notices from the Source form of the Work,
+          excluding those notices that do not pertain to any part of
+          the Derivative Works; and
+
+      (d) If the Work includes a "NOTICE" text file as part of its
+          distribution, then any Derivative Works that You distribute must
+          include a readable copy of the attribution notices contained
+          within such NOTICE file, excluding those notices that do not
+          pertain to any part of the Derivative Works, in at least one
+          of the following places: within a NOTICE text file distributed
+          as part of the Derivative Works; within the Source form or
+          documentation, if provided along with the Derivative Works; or,
+          within a display generated by the Derivative Works, if and
+          wherever such third-party notices normally appear. The contents
+          of the NOTICE file are for informational purposes only and
+          do not modify the License. You may add Your own attribution
+          notices within Derivative Works that You distribute, alongside
+          or as an addendum to the NOTICE text from the Work, provided
+          that such additional attribution notices cannot be construed
+          as modifying the License.
+
+      You may add Your own copyright statement to Your modifications and
+      may provide additional or different license terms and conditions
+      for use, reproduction, or distribution of Your modifications, or
+      for any such Derivative Works as a whole, provided Your use,
+      reproduction, and distribution of the Work otherwise complies with
+      the conditions stated in this License.
+
+   5. Submission of Contributions. Unless You explicitly state otherwise,
+      any Contribution intentionally submitted for inclusion in the Work
+      by You to the Licensor shall be under the terms and conditions of
+      this License, without any additional terms or conditions.
+      Notwithstanding the above, nothing herein shall supersede or modify
+      the terms of any separate license agreement you may have executed
+      with Licensor regarding such Contributions.
+
+   6. Trademarks. This License does not grant permission to use the trade
+      names, trademarks, service marks, or product names of the Licensor,
+      except as required for reasonable and customary use in describing the
+      origin of the Work and reproducing the content of the NOTICE file.
+
+   7. Disclaimer of Warranty. Unless required by applicable law or
+      agreed to in writing, Licensor provides the Work (and each
+      Contributor provides its Contributions) on an "AS IS" BASIS,
+      WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or
+      implied, including, without limitation, any warranties or conditions
+      of TITLE, NON-INFRINGEMENT, MERCHANTABILITY, or FITNESS FOR A
+      PARTICULAR PURPOSE. You are solely responsible for determining the
+      appropriateness of using or redistributing the Work and assume any
+      risks associated with Your exercise of permissions under this License.
+
+   8. Limitation of Liability. In no event and under no legal theory,
+      whether in tort (including negligence), contract, or otherwise,
+      unless required by applicable law (such as deliberate and grossly
+      negligent acts) or agreed to in writing, shall any Contributor be
+      liable to You for damages, including any direct, indirect, special,
+      incidental, or consequential damages of any character arising as a
+      result of this License or out of the use or inability to use the
+      Work (including but not limited to damages for loss of goodwill,
+      work stoppage, computer failure or malfunction, or any and all
+      other commercial damages or losses), even if such Contributor
+      has been advised of the possibility of such damages.
+
+   9. Accepting Warranty or Additional Liability. While redistributing
+      the Work or Derivative Works thereof, You may choose to offer,
+      and charge a fee for, acceptance of support, warranty, indemnity,
+      or other liability obligations and/or rights consistent with this
+      License. However, in accepting such obligations, You may act only
+      on Your own behalf and on Your sole responsibility, not on behalf
+      of any other Contributor, and only if You agree to indemnify,
+      defend, and hold each Contributor harmless for any liability
+      incurred by, or claims asserted against, such Contributor by reason
+      of your accepting any such warranty or additional liability.
+
+   END OF TERMS AND CONDITIONS
+
+   APPENDIX: How to apply the Apache License to your work.
+
+      To apply the Apache License to your work, attach the following
+      boilerplate notice, with the fields enclosed by brackets "[]"
+      replaced with your own identifying information. (Don't include
+      the brackets!)  The text should be enclosed in the appropriate
+      comment syntax for the file format. We also recommend that a
+      file or class name and description of purpose be included on the
+      same "printed page" as the copyright notice for easier
+      identification within third-party archives.
+
+   Copyright [2019] [Division of Medical Image Computing, German Cancer Research Center (DKFZ), Heidelberg, Germany]
+
+   Licensed under the Apache License, Version 2.0 (the "License");
+   you may not use this file except in compliance with the License.
+   You may obtain a copy of the License at
+
+       http://www.apache.org/licenses/LICENSE-2.0
+
+   Unless required by applicable law or agreed to in writing, software
+   distributed under the License is distributed on an "AS IS" BASIS,
+   WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+   See the License for the specific language governing permissions and
+   limitations under the License.

nnUNet/documentation/benchmarking.md

@@ -0,0 +1,115 @@
+# nnU-Netv2 benchmarks
+
+Does your system run like it should? Is your epoch time longer than expected? What epoch times should you expect?
+
+Look no further for we have the solution here!
+
+## What does the nnU-netv2 benchmark do?
+
+nnU-Net's benchmark trains models for 5 epochs. At the end, the fastest epoch will 
+be noted down, along with the GPU name, torch version and cudnn version. You can find the benchmark output in the 
+corresponding nnUNet_results subfolder (see example below). Don't worry, we also provide scripts to collect your 
+results. Or you just start a benchmark and look at the console output. Everything is possible. Nothing is forbidden.
+
+The benchmark implementation revolves around two trainers:
+- `nnUNetTrainerBenchmark_5epochs` runs a regular training for 5 epochs. When completed, writes a .json file with the fastest 
+epoch time as well as the GPU used and the torch and cudnn versions. Useful for speed testing the entire pipeline 
+(data loading, augmentation, GPU training)
+- `nnUNetTrainerBenchmark_5epochs_noDataLoading` is the same, but it doesn't do any data loading or augmentation. It 
+just presents dummy arrays to the GPU. Useful for checking pure GPU speed.
+
+## How to run the nnU-Netv2 benchmark?
+It's quite simple, actually. It looks just like a regular nnU-Net training.
+
+We provide reference numbers for some of the Medical Segmentation Decathlon datasets because they are easily 
+accessible: [download here](https://drive.google.com/drive/folders/1HqEgzS8BV2c7xYNrZdEAnrHk7osJJ--2). If it needs to be 
+quick and dirty, focus on Tasks 2 and 4. Download and extract the data and convert them to the nnU-Net format with 
+`nnUNetv2_convert_MSD_dataset`. 
+Run `nnUNetv2_plan_and_preprocess` for them.
+
+Then, for each dataset, run the following commands (only one per GPU! Or one after the other):
+
+```bash
+nnUNetv2_train DATSET_ID 2d 0 -tr nnUNetTrainerBenchmark_5epochs
+nnUNetv2_train DATSET_ID 3d_fullres 0 -tr nnUNetTrainerBenchmark_5epochs
+nnUNetv2_train DATSET_ID 2d 0 -tr nnUNetTrainerBenchmark_5epochs_noDataLoading
+nnUNetv2_train DATSET_ID 3d_fullres 0 -tr nnUNetTrainerBenchmark_5epochs_noDataLoading
+```
+
+If you want to inspect the outcome manually, check (for example!) your 
+`nnUNet_results/DATASET_NAME/nnUNetTrainerBenchmark_5epochs__nnUNetPlans__3d_fullres/fold_0/` folder for the `benchmark_result.json` file.
+
+Note that there can be multiple entries in this file if the benchmark was run on different GPU types, torch versions or cudnn versions!
+
+If you want to summarize your results like we did in our [results](#results), check the 
+[summary script](../nnunetv2/batch_running/benchmarking/summarize_benchmark_results.py). Here you need to change the 
+torch version, cudnn version and dataset you want to summarize, then execute the script. You can find the exact 
+values you need to put there in one of your `benchmark_result.json` files.
+
+## Results
+We have tested a variety of GPUs and summarized the results in a 
+[spreadsheet](https://docs.google.com/spreadsheets/d/12Cvt_gr8XU2qWaE0XJk5jJlxMEESPxyqW0CWbQhTNNY/edit?usp=sharing). 
+Note that you can select the torch and cudnn versions at the bottom! There may be comments in this spreadsheet. Read them!
+
+## Result interpretation
+
+Results are shown as epoch time in seconds. Lower is better (duh). Epoch times can fluctuate between runs, so as 
+long as you are within like 5-10% of the numbers we report, everything should be dandy. 
+
+If not, here is how you can try to find the culprit!
+
+The first thing to do is to compare the performance between the `nnUNetTrainerBenchmark_5epochs_noDataLoading` and 
+`nnUNetTrainerBenchmark_5epochs` trainers. If the difference is about the same as we report in our spreadsheet, but 
+both your numbers are worse, the problem is with your GPU:
+
+- Are you certain you compare the correct GPU? (duh)
+- If yes, then you might want to install PyTorch in a different way. Never `pip install torch`! Go to the
+[PyTorch installation](https://pytorch.org/get-started/locally/) page, select the most recent cuda version your 
+system supports and only then copy and execute the correct command! Either pip or conda should work
+- If the problem is still not fixed, we recommend you try 
+[compiling pytorch from source](https://github.com/pytorch/pytorch#from-source). It's more difficult but that's 
+how we roll here at the DKFZ (at least the cool kids here).
+- Another thing to consider is to try exactly the same torch + cudnn version as we did in our spreadsheet. 
+Sometimes newer versions can actually degrade performance and there might be bugs from time to time. Older versions 
+are also often a lot slower!
+- Finally, some very basic things that could impact your GPU performance: 
+  - Is the GPU cooled adequately? Check the temperature with `nvidia-smi`. Hot GPUs throttle performance in order to not self-destruct
+  - Is your OS using the GPU for displaying your desktop at the same time? If so then you can expect a performance 
+  penalty (I dunno like 10% !?). That's expected and OK.
+  - Are other users using the GPU as well?
+
+
+If you see a large performance difference between `nnUNetTrainerBenchmark_5epochs_noDataLoading` (fast) and 
+`nnUNetTrainerBenchmark_5epochs` (slow) then the problem might be related to data loading and augmentation. As a 
+reminder, nnU-net does not use pre-augmented images (offline augmentation) but instead generates augmented training 
+samples on the fly during training (no, you cannot switch it to offline). This requires that your system can do partial 
+reads of the image files fast enough (SSD storage required!) and that your CPU is powerful enough to run the augmentations.
+
+Check the following:
+
+- [CPU bottleneck] How many CPU threads are running during the training? nnU-Net uses 12 processes for data augmentation by default. 
+If you see those 12 running constantly during training, consider increasing the number of processes used for data 
+augmentation (provided there is headroom on your CPU!). Increase the number until you see less active workers than 
+you configured (or just set the number to 32 and forget about it). You can do so by setting the `nnUNet_n_proc_DA` 
+environment variable (Linux: `export nnUNet_n_proc_DA=24`). Read [here](set_environment_variables.md) on how to do this.
+If your CPU does not support more processes (setting more processes than your CPU has threads makes 
+no sense!) you are out of luck and in desperate need of a system upgrade!
+- [I/O bottleneck] If you don't see 12 (or nnUNet_n_proc_DA if you set it) processes running but your training times 
+are still slow then open up `top` (sorry, Windows users. I don't know how to do this on Windows) and look at the value 
+left of 'wa' in the row that begins 
+with '%Cpu (s)'. If this is >1.0 (arbitrarily set threshold here, essentially look for unusually high 'wa'. In a 
+healthy training 'wa' will be almost 0) then your storage cannot keep up with data loading. Make sure to set 
+nnUNet_preprocessed to a folder that is located on an SSD. nvme is preferred over SATA. PCIe3 is enough. 3000MB/s 
+sequential read recommended.
+- [funky stuff] Sometimes there is funky stuff going on, especially when batch sizes are large, files are small and 
+patch sizes are small as well. As part of the data loading process, nnU-Net needs to open and close a file for each 
+training sample. Now imagine a dataset like Dataset004_Hippocampus where for the 2d config we have a batch size of 
+366 and we run 250 iterations in <10s on an A100. That's a lotta files per second (366 * 250 / 10 = 9150 files per second). 
+Oof. If the files are on some network drive (even if it's nvme) then (probably) good night. The good news: nnU-Net
+has got you covered: add `export nnUNet_keep_files_open=True` to your .bashrc and the problem goes away. The neat 
+part: it causes new problems if you are not allowed to have enough open files. You may have to increase the number 
+of allowed open files. `ulimit -n` gives your current limit (Linux only). It should not be something like 1024. 
+Increasing that to 65535 works well for me. See here for how to change these limits: 
+[Link](https://kupczynski.info/posts/ubuntu-18-10-ulimits/) 
+(works for Ubuntu 18, google for your OS!).
+