Geometry flip

.github/workflows/conda.yml

@@ -26,7 +26,7 @@ jobs:
     steps:
     - if: runner.os == 'windows'
       name: Config pagefile (Windows only)
-      uses: al-cheb/configure-pagefile-action@v1.3
+      uses: al-cheb/configure-pagefile-action@v1.4
       with:
         minimum-size: 8GB
         maximum-size: 16GB

.github/workflows/cron.yml

@@ -19,18 +19,18 @@ jobs:
           - "PTLATEST+CUDA121"
         include:
           # https://docs.nvidia.com/deeplearning/frameworks/pytorch-release-notes
-          - environment: PT191+CUDA113
-            pytorch: "torch==1.9.1 torchvision==0.10.1 --extra-index-url https://download.pytorch.org/whl/cu113"
-            base: "nvcr.io/nvidia/pytorch:21.06-py3"  # CUDA 11.3
           - environment: PT110+CUDA113
             pytorch: "torch==1.10.2 torchvision==0.11.3 --extra-index-url https://download.pytorch.org/whl/cu113"
             base: "nvcr.io/nvidia/pytorch:21.06-py3"  # CUDA 11.3
           - environment: PT113+CUDA113
             pytorch: "torch==1.13.1 torchvision==0.14.1 --extra-index-url https://download.pytorch.org/whl/cu113"
             base: "nvcr.io/nvidia/pytorch:21.06-py3"  # CUDA 11.3
-          - environment: PTLATEST+CUDA121
-            pytorch: "-U torch torchvision --extra-index-url https://download.pytorch.org/whl/cu118"
+          - environment: PT113+CUDA122
+            pytorch: "torch==1.13.1 torchvision==0.14.1 --extra-index-url https://download.pytorch.org/whl/cu121"
             base: "nvcr.io/nvidia/pytorch:23.08-py3"  # CUDA 12.2
+          - environment: PTLATEST+CUDA124
+            pytorch: "-U torch torchvision --extra-index-url https://download.pytorch.org/whl/cu121"
+            base: "nvcr.io/nvidia/pytorch:24.03-py3"  # CUDA 12.4
     container:
       image: ${{ matrix.base }}
       options: "--gpus all"
@@ -76,7 +76,7 @@ jobs:
     if: github.repository == 'Project-MONAI/MONAI'
     strategy:
       matrix:
-        container: ["pytorch:22.10", "pytorch:23.08"]
+        container: ["pytorch:23.08", "pytorch:24.03"]
     container:
       image: nvcr.io/nvidia/${{ matrix.container }}-py3  # testing with the latest pytorch base image
       options: "--gpus all"
@@ -121,7 +121,7 @@ jobs:
     if: github.repository == 'Project-MONAI/MONAI'
     strategy:
       matrix:
-        container: ["pytorch:23.08"]
+        container: ["pytorch:24.03"]
     container:
       image: nvcr.io/nvidia/${{ matrix.container }}-py3  # testing with the latest pytorch base image
       options: "--gpus all"
@@ -221,7 +221,7 @@ jobs:
     if: github.repository == 'Project-MONAI/MONAI'
     needs: cron-gpu  # so that monai itself is verified first
     container:
-      image: nvcr.io/nvidia/pytorch:23.08-py3  # testing with the latest pytorch base image
+      image: nvcr.io/nvidia/pytorch:24.03-py3  # testing with the latest pytorch base image
       options: "--gpus all --ipc=host"
     runs-on: [self-hosted, linux, x64, integration]
     steps:

.github/workflows/pythonapp-gpu.yml

@@ -29,10 +29,6 @@ jobs:
           - "PT210+CUDA121DOCKER"
         include:
           # https://docs.nvidia.com/deeplearning/frameworks/pytorch-release-notes
-          - environment: PT19+CUDA114DOCKER
-            # 21.10: 1.10.0a0+0aef44c
-            pytorch: "-h"  # we explicitly set pytorch to -h to avoid pip install error
-            base: "nvcr.io/nvidia/pytorch:21.10-py3"
           - environment: PT110+CUDA111
             pytorch: "torch==1.10.2 torchvision==0.11.3 --extra-index-url https://download.pytorch.org/whl/cu111"
             base: "nvcr.io/nvidia/cuda:11.1.1-devel-ubuntu18.04"
@@ -47,6 +43,10 @@ jobs:
             # 23.08: 2.1.0a0+29c30b1
             pytorch: "-h"  # we explicitly set pytorch to -h to avoid pip install error
             base: "nvcr.io/nvidia/pytorch:23.08-py3"
+          - environment: PT210+CUDA121DOCKER
+            # 24.03: 2.3.0a0+40ec155e58.nv24.3
+            pytorch: "-h"  # we explicitly set pytorch to -h to avoid pip install error
+            base: "nvcr.io/nvidia/pytorch:24.03-py3"
     container:
       image: ${{ matrix.base }}
       options: --gpus all --env NVIDIA_DISABLE_REQUIRE=true  # workaround for unsatisfied condition: cuda>=11.6

.github/workflows/pythonapp.yml

@@ -62,7 +62,7 @@ jobs:
     steps:
     - if: runner.os == 'windows'
       name: Config pagefile (Windows only)
-      uses: al-cheb/configure-pagefile-action@v1.3
+      uses: al-cheb/configure-pagefile-action@v1.4
       with:
         minimum-size: 8GB
         maximum-size: 16GB

Dockerfile

@@ -11,11 +11,15 @@
 
 # To build with a different base image
 # please run `docker build` using the `--build-arg PYTORCH_IMAGE=...` flag.
-ARG PYTORCH_IMAGE=nvcr.io/nvidia/pytorch:23.08-py3
+ARG PYTORCH_IMAGE=nvcr.io/nvidia/pytorch:24.03-py3
 FROM ${PYTORCH_IMAGE}
 
 LABEL maintainer="monai.contact@gmail.com"
 
+# TODO: remark for issue [revise the dockerfile](https://github.com/zarr-developers/numcodecs/issues/431)
+WORKDIR /opt
+RUN git clone --recursive https://github.com/zarr-developers/numcodecs.git && pip wheel numcodecs
+
 WORKDIR /opt/monai
 
 # install full deps

docs/source/losses.rst

@@ -73,6 +73,11 @@ Segmentation Losses
 .. autoclass:: ContrastiveLoss
     :members:
 
+`BarlowTwinsLoss`
+~~~~~~~~~~~~~~~~~
+.. autoclass:: BarlowTwinsLoss
+    :members:
+
 `HausdorffDTLoss`
 ~~~~~~~~~~~~~~~~~
 .. autoclass:: HausdorffDTLoss
@@ -134,6 +139,11 @@ Reconstruction Losses
 .. autoclass:: JukeboxLoss
     :members:
 
+`SURELoss`
+~~~~~~~~~~
+.. autoclass:: SURELoss
+    :members:
+
 
 Loss Wrappers
 -------------

docs/source/networks.rst

@@ -408,6 +408,11 @@ Layers
 .. autoclass:: LLTM
     :members:
 
+`ConjugateGradient`
+~~~~~~~~~~~~~~~~~~~
+.. autoclass:: ConjugateGradient
+    :members:
+
 `Utilities`
 ~~~~~~~~~~~
 .. automodule:: monai.networks.layers.convutils

docs/source/transforms.rst

@@ -309,6 +309,12 @@ Intensity
     :members:
     :special-members: __call__
 
+`ClipIntensityPercentiles`
+""""""""""""""""""""""""""
+.. autoclass:: ClipIntensityPercentiles
+    :members:
+    :special-members: __call__
+
 `RandScaleIntensity`
 """"""""""""""""""""
 .. image:: https://raw.githubusercontent.com/Project-MONAI/DocImages/main/transforms/RandScaleIntensity.png
@@ -661,6 +667,27 @@ Post-processing
     :members:
     :special-members: __call__
 
+Regularization
+^^^^^^^^^^^^^^
+
+`CutMix`
+""""""""
+.. autoclass:: CutMix
+    :members:
+    :special-members: __call__
+
+`CutOut`
+""""""""
+.. autoclass:: CutOut
+    :members:
+    :special-members: __call__
+
+`MixUp`
+"""""""
+.. autoclass:: MixUp
+    :members:
+    :special-members: __call__
+
 Signal
 ^^^^^^^
 
@@ -1384,6 +1411,12 @@ Intensity (Dict)
     :members:
     :special-members: __call__
 
+`ClipIntensityPercentilesd`
+"""""""""""""""""""""""""""
+.. autoclass:: ClipIntensityPercentilesd
+    :members:
+    :special-members: __call__
+
 `RandScaleIntensityd`
 """""""""""""""""""""
 .. image:: https://raw.githubusercontent.com/Project-MONAI/DocImages/main/transforms/RandScaleIntensityd.png
@@ -1707,6 +1740,27 @@ Post-processing (Dict)
   :members:
   :special-members: __call__
 
+Regularization (Dict)
+^^^^^^^^^^^^^^^^^^^^^
+
+`CutMixd`
+"""""""""
+.. autoclass:: CutMixd
+    :members:
+    :special-members: __call__
+
+`CutOutd`
+"""""""""
+.. autoclass:: CutOutd
+    :members:
+    :special-members: __call__
+
+`MixUpd`
+""""""""
+.. autoclass:: MixUpd
+    :members:
+    :special-members: __call__
+
 Signal (Dict)
 ^^^^^^^^^^^^^
 

docs/source/transforms_idx.rst

@@ -74,6 +74,16 @@ Post-processing
    post.array
    post.dictionary
 
+Regularization
+^^^^^^^^^^^^^^
+
+.. autosummary::
+   :toctree: _gen
+   :nosignatures:
+
+   regularization.array
+   regularization.dictionary
+
 Signal
 ^^^^^^
 

monai/__init__.py

@@ -83,6 +83,11 @@
     from .utils.tf32 import detect_default_tf32
 
     detect_default_tf32()
+    import torch
+
+    # workaround related to https://github.com/Project-MONAI/MONAI/issues/7575
+    if hasattr(torch.cuda.device_count, "cache_clear"):
+        torch.cuda.device_count.cache_clear()
 except BaseException:
     from .utils.misc import MONAIEnvVars
 

monai/apps/auto3dseg/hpo_gen.py

@@ -193,7 +193,9 @@ def generate(self, output_folder: str = ".") -> None:
         self.obj_filename = os.path.join(write_path, "algo_object.pkl")
 
         if isinstance(self.algo, BundleAlgo):
-            self.algo.export_to_disk(output_folder, task_prefix + task_id, fill_with_datastats=False)
+            self.algo.export_to_disk(
+                output_folder, task_prefix + task_id, bundle_root=write_path, fill_with_datastats=False
+            )
         else:
             ConfigParser.export_config_file(self.params, write_path)
             logger.info(write_path)

monai/apps/nnunet/nnunetv2_runner.py

@@ -22,6 +22,7 @@
 from monai.apps.nnunet.utils import analyze_data, create_new_data_copy, create_new_dataset_json
 from monai.bundle import ConfigParser
 from monai.utils import ensure_tuple, optional_import
+from monai.utils.misc import run_cmd
 
 load_pickle, _ = optional_import("batchgenerators.utilities.file_and_folder_operations", name="load_pickle")
 join, _ = optional_import("batchgenerators.utilities.file_and_folder_operations", name="join")
@@ -495,65 +496,64 @@ def train_single_model(self, config: Any, fold: int, gpu_id: tuple | list | int
             fold: fold of the 5-fold cross-validation. Should be an int between 0 and 4.
             gpu_id: an integer to select the device to use, or a tuple/list of GPU device indices used for multi-GPU
                 training (e.g., (0,1)). Default: 0.
-        from nnunetv2.run.run_training import run_training
             kwargs: this optional parameter allows you to specify additional arguments in
-                ``nnunetv2.run.run_training.run_training``. Currently supported args are
-                    - plans_identifier: custom plans identifier. Default: "nnUNetPlans".
-                    - pretrained_weights: path to nnU-Net checkpoint file to be used as pretrained model. Will only be
-                        used when actually training. Beta. Use with caution. Default: False.
-                    - use_compressed_data: True to use compressed data for training. Reading compressed data is much
-                        more CPU and (potentially) RAM intensive and should only be used if you know what you are
-                        doing. Default: False.
-                    - continue_training: continue training from latest checkpoint. Default: False.
-                    - only_run_validation: True to run the validation only. Requires training to have finished.
-                        Default: False.
-                    - disable_checkpointing: True to disable checkpointing. Ideal for testing things out and you
-                        don't want to flood your hard drive with checkpoints. Default: False.
+                ``nnunetv2.run.run_training.run_training_entry``.
+
+                Currently supported args are:
+
+                - p: custom plans identifier. Default: "nnUNetPlans".
+                - pretrained_weights: path to nnU-Net checkpoint file to be used as pretrained model. Will only be
+                    used when actually training. Beta. Use with caution. Default: False.
+                - use_compressed: True to use compressed data for training. Reading compressed data is much
+                    more CPU and (potentially) RAM intensive and should only be used if you know what you are
+                    doing. Default: False.
+                - c: continue training from latest checkpoint. Default: False.
+                - val: True to run the validation only. Requires training to have finished.
+                    Default: False.
+                - disable_checkpointing: True to disable checkpointing. Ideal for testing things out and you
+                    don't want to flood your hard drive with checkpoints. Default: False.
         """
         if "num_gpus" in kwargs:
             kwargs.pop("num_gpus")
             logger.warning("please use gpu_id to set the GPUs to use")
 
-        if "trainer_class_name" in kwargs:
-            kwargs.pop("trainer_class_name")
+        if "tr" in kwargs:
+            kwargs.pop("tr")
             logger.warning("please specify the `trainer_class_name` in the __init__ of `nnUNetV2Runner`.")
 
-        if "export_validation_probabilities" in kwargs:
-            kwargs.pop("export_validation_probabilities")
+        if "npz" in kwargs:
+            kwargs.pop("npz")
             logger.warning("please specify the `export_validation_probabilities` in the __init__ of `nnUNetV2Runner`.")
 
+        cmd = self.train_single_model_command(config, fold, gpu_id, kwargs)
+        run_cmd(cmd, shell=True)
+
+    def train_single_model_command(self, config, fold, gpu_id, kwargs):
         if isinstance(gpu_id, (tuple, list)):
             if len(gpu_id) > 1:
                 gpu_ids_str = ""
                 for _i in range(len(gpu_id)):
                     gpu_ids_str += f"{gpu_id[_i]},"
-                os.environ["CUDA_VISIBLE_DEVICES"] = gpu_ids_str[:-1]
+                device_setting = f"CUDA_VISIBLE_DEVICES={gpu_ids_str[:-1]}"
             else:
-                os.environ["CUDA_VISIBLE_DEVICES"] = f"{gpu_id[0]}"
-        else:
-            os.environ["CUDA_VISIBLE_DEVICES"] = f"{gpu_id}"
-
-        from nnunetv2.run.run_training import run_training
-
-        if isinstance(gpu_id, int) or len(gpu_id) == 1:
-            run_training(
-                dataset_name_or_id=self.dataset_name_or_id,
-                configuration=config,
-                fold=fold,
-                trainer_class_name=self.trainer_class_name,
-                export_validation_probabilities=self.export_validation_probabilities,
-                **kwargs,
-            )
+                device_setting = f"CUDA_VISIBLE_DEVICES={gpu_id[0]}"
         else:
-            run_training(
-                dataset_name_or_id=self.dataset_name_or_id,
-                configuration=config,
-                fold=fold,
-                num_gpus=len(gpu_id),
-                trainer_class_name=self.trainer_class_name,
-                export_validation_probabilities=self.export_validation_probabilities,
-                **kwargs,
-            )
+            device_setting = f"CUDA_VISIBLE_DEVICES={gpu_id}"
+        num_gpus = 1 if isinstance(gpu_id, int) or len(gpu_id) == 1 else len(gpu_id)
+
+        cmd = (
+            f"{device_setting} nnUNetv2_train "
+            + f"{self.dataset_name_or_id} {config} {fold} "
+            + f"-tr {self.trainer_class_name} -num_gpus {num_gpus}"
+        )
+        if self.export_validation_probabilities:
+            cmd += " --npz"
+        for _key, _value in kwargs.items():
+            if _key == "p" or _key == "pretrained_weights":
+                cmd += f" -{_key} {_value}"
+            else:
+                cmd += f" --{_key} {_value}"
+        return cmd
 
     def train(
         self,
@@ -637,15 +637,7 @@ def train_parallel_cmd(
                 if _config in ensure_tuple(configs):
                     for _i in range(self.num_folds):
                         the_device = gpu_id_for_all[_index % n_devices]  # type: ignore
-                        cmd = (
-                            "python -m monai.apps.nnunet nnUNetV2Runner train_single_model "
-                            + f"--input_config '{self.input_config_or_dict}' --work_dir '{self.work_dir}' "
-                            + f"--config '{_config}' --fold {_i} --gpu_id {the_device} "
-                            + f"--trainer_class_name {self.trainer_class_name} "
-                            + f"--export_validation_probabilities {self.export_validation_probabilities}"
-                        )
-                        for _key, _value in kwargs.items():
-                            cmd += f" --{_key} {_value}"
+                        cmd = self.train_single_model_command(_config, _i, the_device, kwargs)
                         all_cmds[-1][the_device].append(cmd)
                         _index += 1
         return all_cmds