Skip to content

Latest commit

 

History

History
129 lines (102 loc) · 5.63 KB

AutoPETII.md

File metadata and controls

129 lines (102 loc) · 5.63 KB

Look Ma, no code: fine tuning nnU-Net for the AutoPET II challenge by only adjusting its JSON plans

Please cite our paper :-*

COMING SOON

Intro

See the Challenge Website for details on the challenge.

Our solution to this challenge rewuires no code changes at all. All we do is optimize nnU-Net's hyperparameters (architecture, batch size, patch size) through modifying the nnUNetplans.json file.

Prerequisites

Use the latest pytorch version!

We recommend you use the latest nnU-Net version as well! We ran our trainings with commit 913705f which you can try in case something doesn't work as expected: pip install git+https://github.com/MIC-DKFZ/nnUNet.git@913705f

How to reproduce our trainings

Download and convert the data

  1. Download and extract the AutoPET II dataset
  2. Convert it to nnU-Net format by running python nnunetv2/dataset_conversion/Dataset221_AutoPETII_2023.py FOLDER where folder is the extracted AutoPET II dataset.

Experiment planning and preprocessing

We deviate a little from the standard nnU-Net procedure because all our experiments are based on just the 3d_fullres configuration

Run the following commands:

  • nnUNetv2_extract_fingerprint -d 221 extracts the dataset fingerprint
  • nnUNetv2_plan_experiment -d 221 does the planning for the plain unet
  • nnUNetv2_plan_experiment -d 221 -pl ResEncUNetPlanner does the planning for the residual encoder unet
  • nnUNetv2_preprocess -d 221 -c 3d_fullres runs all the preprocessing we need

Modification of plans files

Please read the information on how to modify plans files first!!!

It is easier to have everything in one plans file, so the first thing we do is transfer the ResEnc UNet to the default plans file. We use the configuration inheritance feature of nnU-Net to make it use the same data as the 3d_fullres configuration. Add the following to the 'configurations' dict in 'nnUNetPlans.json':

        "3d_fullres_resenc": {
            "inherits_from": "3d_fullres",
            "network_arch_class_name": "ResidualEncoderUNet",
            "n_conv_per_stage_encoder": [
                1,
                3,
                4,
                6,
                6,
                6
            ],
            "n_conv_per_stage_decoder": [
                1,
                1,
                1,
                1,
                1
            ]
        },

(these values are basically just copied from the 'nnUNetResEncUNetPlans.json' file! With everything redundant being omitted thanks to inheritance from 3d_fullres)

Now we crank up the patch and batch sizes. Add the following configurations:

        "3d_fullres_resenc_bs80": {
            "inherits_from": "3d_fullres_resenc",
            "batch_size": 80
            },
        "3d_fullres_resenc_192x192x192_b24": {
            "inherits_from": "3d_fullres_resenc",
            "patch_size": [
                192,
                192,
                192
            ],
            "batch_size": 24
        }

Save the file (and check for potential Syntax Errors!)

Run trainings

Training each model requires 8 Nvidia A100 40GB GPUs. Expect training to run for 5-7 days. You'll need a really good CPU to handle the data augmentation! 128C/256T are a must! If you have less threads available, scale down nnUNet_n_proc_DA accordingly.

nnUNet_compile=T nnUNet_n_proc_DA=28 nnUNetv2_train 221 3d_fullres_resenc_bs80 0 -num_gpus 8
nnUNet_compile=T nnUNet_n_proc_DA=28 nnUNetv2_train 221 3d_fullres_resenc_bs80 1 -num_gpus 8
nnUNet_compile=T nnUNet_n_proc_DA=28 nnUNetv2_train 221 3d_fullres_resenc_bs80 2 -num_gpus 8
nnUNet_compile=T nnUNet_n_proc_DA=28 nnUNetv2_train 221 3d_fullres_resenc_bs80 3 -num_gpus 8
nnUNet_compile=T nnUNet_n_proc_DA=28 nnUNetv2_train 221 3d_fullres_resenc_bs80 4 -num_gpus 8

nnUNet_compile=T nnUNet_n_proc_DA=28 nnUNetv2_train 221 3d_fullres_resenc_192x192x192_b24 0 -num_gpus 8
nnUNet_compile=T nnUNet_n_proc_DA=28 nnUNetv2_train 221 3d_fullres_resenc_192x192x192_b24 1 -num_gpus 8
nnUNet_compile=T nnUNet_n_proc_DA=28 nnUNetv2_train 221 3d_fullres_resenc_192x192x192_b24 2 -num_gpus 8
nnUNet_compile=T nnUNet_n_proc_DA=28 nnUNetv2_train 221 3d_fullres_resenc_192x192x192_b24 3 -num_gpus 8
nnUNet_compile=T nnUNet_n_proc_DA=28 nnUNetv2_train 221 3d_fullres_resenc_192x192x192_b24 4 -num_gpus 8

Done!

(We also provide pretrained weights in case you don't want to invest the GPU resources, see below)

How to make predictions with pretrained weights

Our final model is an ensemble of two configurations:

  • ResEnc UNet with batch size 80
  • ResEnc UNet with patch size 192x192x192 and batch size 24

To run inference with these models, do the following:

  1. Download the pretrained model weights from Zenodo
  2. Install both .zip files using nnUNetv2_install_pretrained_model_from_zip
  3. Make sure
  4. Now you can run inference on new cases with nnUNetv2_predict:
    • nnUNetv2_predict -i INPUT -o OUTPUT1 -d 221 -c 3d_fullres_resenc_bs80 -f 0 1 2 3 4 -step_size 0.6 --save_probabilities
    • nnUNetv2_predict -i INPUT -o OUTPUT2 -d 221 -c 3d_fullres_resenc_192x192x192_b24 -f 0 1 2 3 4 --save_probabilities
    • nnUNetv2_ensemble -i OUTPUT1 OUTPUT2 -o OUTPUT_ENSEMBLE

Note that our inference Docker omitted TTA via mirroring along the axial direction during prediction (only sagittal + coronal mirroring). This was done to keep the inference time below 10 minutes per image on a T4 GPU (we actually never tested whether we could have left this enabled). Just leave it on! You can also leave the step_size at default for the 3d_fullres_resenc_bs80.