GANDLF 0.1.3.dev20250319__py3-none-any.whl → 0.1.4.dev20250503__py3-none-any.whl

GANDLF/utils/imaging.py

@@ -5,6 +5,7 @@ import numpy as np
 import SimpleITK as sitk
 import torchio
 import cv2
+from skimage.io import imsave
 
 from .generic import get_filename_extension_sanitized
 
@@ -230,7 +231,7 @@ def write_training_patches(subject: torchio.Subject, params: dict) -> None:
     ext = get_filename_extension_sanitized(subject["path_to_metadata"][0])
     for key in params["channel_keys"]:
         img_to_write = torchio.ScalarImage(
-            tensor=subject[key][torchio.DATA][0], affine=subject[key]["affine"][0]
+            tensor=subject[key][torchio.DATA][0].cpu(), affine=subject[key]["affine"][0]
         ).as_sitk()
         sitk.WriteImage(
             img_to_write,
@@ -239,7 +240,7 @@ def write_training_patches(subject: torchio.Subject, params: dict) -> None:
 
     if params["label_keys"] is not None:
         img_to_write = torchio.ScalarImage(
-            tensor=subject[params["label_keys"][0]][torchio.DATA][0],
+            tensor=subject[params["label_keys"][0]][torchio.DATA][0].cpu(),
             affine=subject[key]["affine"][0],
         ).as_sitk()
         sitk.WriteImage(
@@ -285,3 +286,121 @@ def applyCustomColorMap(im_gray: np.ndarray) -> np.ndarray:
     lut[:, 0, 1] = np.zeros((256)).tolist()
     lut[:, 0, 2] = np.arange(0, 256, 1).tolist()
     return cv2.LUT(img_bgr, lut)
+
+
+class MapSaver:
+    ALLOWED_COLORMAPS = ["jet", "turbo", "angi"]
+
+    def __init__(
+        self,
+        num_classes: int,
+        slide_level: int,
+        blending_alpha: float,
+        level_width: int,
+        level_height: int,
+    ):
+        self.num_classes = num_classes
+        self.slide_level = slide_level
+        self.blending_alpha = blending_alpha
+        self.level_width = level_width
+        self.level_height = level_height
+
+    def save_count_map(self, count_map, save_dir):
+        count_map = np.array(count_map * 255, dtype=np.uint16)
+        save_path = os.path.join(save_dir, "count_map.png")
+        imsave(save_path, count_map)
+
+    def save_probability_and_segmentation_maps(
+        self, probability_map, openslide_image, save_dir
+    ):
+        (
+            heatmaps,
+            segmentation_maps,
+        ) = self._create_heatmaps_and_segmentation_from_probability_map(probability_map)
+        self._save_segmentation_maps(segmentation_maps, save_dir)
+        self._save_heatmaps(heatmaps, save_dir)
+        self._save_heatmaps_overlaid_on_slide(heatmaps, openslide_image, save_dir)
+
+    def _create_heatmaps_and_segmentation_from_probability_map(self, probability_map):
+        heatmaps = {}
+        segmentation_maps = {}
+        for class_id in range(self.num_classes):
+            heatmap_gray = self._create_grey_heatmap(probability_map, class_id)
+            heatmap_turbo = self._apply_colormap(heatmap_gray, "turbo")
+            heatmap_jet = self._apply_colormap(heatmap_gray, "jet")
+            heatmap_angi = self._apply_colormap(heatmap_gray, "angi")
+
+            heatmaps[f"{class_id}_turbo"] = heatmap_turbo
+            heatmaps[f"{class_id}_jet"] = heatmap_jet
+            heatmaps[f"{class_id}_angi"] = heatmap_angi
+
+            segmap = ((probability_map[class_id, ...] > 0.5).astype(np.uint8)) * 255
+            segmentation_maps[class_id] = segmap
+
+        return heatmaps, segmentation_maps
+
+    def _save_segmentation_maps(self, segmentation_maps, save_dir):
+        for class_id, seg_map in segmentation_maps.items():
+            save_path = os.path.join(save_dir, f"seg_map_{class_id}.png")
+            cv2.imwrite(save_path, seg_map)
+
+    def _save_heatmaps(self, heatmaps, save_dir):
+        for class_id_and_colormap, heatmap in heatmaps.items():
+            save_path = os.path.join(
+                save_dir, f"probability_map_{class_id_and_colormap}.png"
+            )
+            cv2.imwrite(save_path, heatmap)
+
+    def _save_heatmaps_overlaid_on_slide(self, heatmaps, openslide_image, save_dir):
+        for class_id_and_colormap, heatmap in heatmaps.items():
+            save_path = os.path.join(
+                save_dir, f"probability_map_blended_{class_id_and_colormap}.png"
+            )
+            blended_heatmap = self._blend_heatmap_with_slide(heatmap, openslide_image)
+            cv2.imwrite(save_path, blended_heatmap)
+
+    def _blend_heatmap_with_slide(self, heatmap, openslide_image):
+        openslide_image_as_array = np.array(
+            openslide_image.read_region(
+                (0, 0), self.slide_level, (self.level_width, self.level_height)
+            ).convert("RGB")
+        )
+        blended_heatmap = cv2.addWeighted(
+            openslide_image_as_array,
+            self.blending_alpha,
+            heatmap,
+            1 - self.blending_alpha,
+            0,
+        )
+        return blended_heatmap
+
+    def _create_grey_heatmap(self, probability_map, class_id):
+        heatmap_gray = np.array(probability_map[class_id, ...] * 255, dtype=np.uint8)
+        return heatmap_gray
+
+    def _apply_colormap(self, heatmap_gray, colormap_type):
+        colormap = colormap_type.lower()
+        assert (
+            colormap in self.ALLOWED_COLORMAPS
+        ), f"Colormap {colormap} not supported. Supported colormaps are {self.ALLOWED_COLORMAPS}"
+
+        if colormap == "jet":
+            return self._apply_jet_colormap(heatmap_gray)
+        elif colormap == "turbo":
+            return self._apply_turbo_colormap(heatmap_gray)
+
+        return self._apply_angi_colormap(heatmap_gray)
+
+    def _apply_jet_colormap(self, heatmap_gray):
+        return cv2.applyColorMap(heatmap_gray, cv2.COLORMAP_JET)
+
+    def _apply_turbo_colormap(self, heatmap_gray):
+        return cv2.applyColorMap(heatmap_gray, cv2.COLORMAP_TURBO)
+
+    def _apply_angi_colormap(self, heatmap_gray):
+        img_bgr = cv2.cvtColor(heatmap_gray.astype(np.uint8), cv2.COLOR_BGR2RGB)
+        lut = np.zeros((256, 1, 3), dtype=np.uint8)
+        lut[:, 0, 0] = np.zeros((256)).tolist()
+        lut[:, 0, 1] = np.zeros((256)).tolist()
+        lut[:, 0, 2] = np.arange(0, 256, 1).tolist()
+        return cv2.LUT(img_bgr, lut)

GANDLF/utils/modelio.py

@@ -20,9 +20,9 @@ model_dict_full = {
 
 model_dict_required = {"model_state_dict": None, "optimizer_state_dict": None}
 
-best_model_path_end = "_best.pth.tar"
-latest_model_path_end = "_latest.pth.tar"
-initial_model_path_end = "_initial.pth.tar"
+BEST_MODEL_PATH_END = "_best.pth.tar"
+LATEST_MODEL_PATH_END = "_latest.pth.tar"
+INITIAL_MODEL_PATH_END = "_initial.pth.tar"
 
 
 def optimize_and_save_model(
@@ -96,10 +96,12 @@ def optimize_and_save_model(
             from openvino.tools.mo import convert_model
             from openvino.runtime import get_version
 
-            openvino_present = False
-            # check for the correct openvino version to prevent inadvertent api breaks
-            if "2023.0.1" in get_version():
-                openvino_present = True
+            # debugging
+            openvino_present = True
+            # openvino_present = False
+            # # check for the correct openvino version to prevent inadvertent api breaks
+            # if "2024.1.0" in get_version():
+            #     openvino_present = True
         except ImportError:
             print("WARNING: OpenVINO is not present.")
 

GANDLF/utils/pred_target_processors.py

@@ -0,0 +1,71 @@
+import torch
+import torch.nn.functional as F
+from abc import ABC, abstractmethod
+from GANDLF.utils.tensor import reverse_one_hot, get_linear_interpolation_mode, one_hot
+
+from typing import Tuple
+
+
+class AbstractPredictionTargetProcessor(ABC):
+    def __init__(self, params: dict):
+        """
+        Interface for classes that perform specific processing on the target and/or prediction tensors.
+        Useful for example for metrics or loss calculations, where some architectures require specific
+        processing of the target and/or prediction tensors before the metric or loss can be calculated.
+        """
+        super().__init__()
+        self.params = params
+
+    @abstractmethod
+    def __call__(
+        self, prediction: torch.Tensor, target: torch.Tensor, *args
+    ) -> Tuple[torch.Tensor, torch.Tensor]:
+        pass
+
+
+class DeepSupervisionPredictionTargetProcessor(AbstractPredictionTargetProcessor):
+    def __init__(self, params: dict):
+        """
+        Processor for deep supervision architectures.
+        """
+        super().__init__(params)
+
+    def __call__(self, prediction: torch.Tensor, target: torch.Tensor, *args):
+        target = one_hot(target, self.params["model"]["class_list"])
+        target_resampled = []
+        target_prev = target.detach()
+        for i, _ in enumerate(prediction):
+            if target_prev[0].shape != prediction[i][0].shape:
+                expected_shape = reverse_one_hot(
+                    prediction[i][0].detach(), self.params["model"]["class_list"]
+                ).shape
+                target_prev = F.interpolate(
+                    target_prev,
+                    size=expected_shape,
+                    mode=get_linear_interpolation_mode(len(expected_shape)),
+                    align_corners=False,
+                )
+            target_resampled.append(target_prev)
+        return prediction, target_resampled
+
+
+class IdentityPredictionTargetProcessor(AbstractPredictionTargetProcessor):
+    def __init__(self, params: dict):
+        """
+        No-op processor that returns the input target and prediction tensors.
+        Used when no processing is needed.
+        """
+        super().__init__(params)
+
+    def __call__(self, prediction: torch.Tensor, target: torch.Tensor, *args):
+        return prediction, target
+
+
+class PredictionTargetProcessorFactory:
+    def __init__(self, params: dict):
+        self.params = params
+
+    def get_prediction_target_processor(self) -> AbstractPredictionTargetProcessor:
+        if "deep" in self.params["model"]["architecture"].lower():
+            return DeepSupervisionPredictionTargetProcessor(self.params)
+        return IdentityPredictionTargetProcessor(self.params)

GANDLF/utils/write_parse.py

@@ -100,7 +100,7 @@ def parseTrainingCSV(
 
     for col in data_full.columns:
         # add appropriate headers to read here, as needed
-        col_lower = col.lower()
+        col_lower = str(col).lower()
         currentHeaderLoc = data_full.columns.get_loc(col)
         if (
             ("channel" in col_lower)

GANDLF/version.py

@@ -1 +1 @@
-__version__ = "0.1.3-dev20250319"
+__version__ = "0.1.4-dev20250503"

{gandlf-0.1.3.dev20250319.dist-info → gandlf-0.1.4.dev20250503.dist-info}/METADATA

@@ -1,6 +1,6 @@
-Metadata-Version: 2.2
+Metadata-Version: 2.4
 Name: GANDLF
-Version: 0.1.3.dev20250319
+Version: 0.1.4.dev20250503
 Summary: PyTorch-based framework that handles segmentation/regression/classification using various DL architectures for medical imaging.
 Author: MLCommons
 Author-email: gandlf@mlcommons.org
@@ -13,13 +13,15 @@ Classifier: Natural Language :: English
 Classifier: Operating System :: OS Independent
 Classifier: Programming Language :: Python :: 3.10
 Classifier: Programming Language :: Python :: 3.11
+Classifier: Programming Language :: Python :: 3.12
 Classifier: Topic :: Scientific/Engineering :: Medical Science Apps.
-Requires-Python: >3.8, <3.12
+Requires-Python: >3.9, <3.13
 Description-Content-Type: text/markdown
 License-File: LICENSE
 Requires-Dist: torch==2.5.0
 Requires-Dist: black==23.11.0
-Requires-Dist: numpy==1.25.0
+Requires-Dist: lightning==2.4.0
+Requires-Dist: numpy==1.26.4
 Requires-Dist: scipy
 Requires-Dist: SimpleITK!=2.0.*
 Requires-Dist: SimpleITK!=2.2.1
@@ -31,9 +33,10 @@ Requires-Dist: scikit-learn>=0.23.2
 Requires-Dist: scikit-image>=0.19.1
 Requires-Dist: setuptools
 Requires-Dist: seaborn
-Requires-Dist: pyyaml==6.0.1
+Requires-Dist: pyyaml
 Requires-Dist: matplotlib
 Requires-Dist: gdown==5.1.0
+Requires-Dist: overrides==7.7.0
 Requires-Dist: pytest
 Requires-Dist: coverage
 Requires-Dist: pytest-cov
@@ -42,18 +45,18 @@ Requires-Dist: medcam
 Requires-Dist: opencv-python
 Requires-Dist: torchmetrics==1.1.2
 Requires-Dist: zarr==2.10.3
+Requires-Dist: numcodecs<0.16.0
 Requires-Dist: pydicom
 Requires-Dist: onnx
 Requires-Dist: torchinfo==1.7.0
-Requires-Dist: segmentation-models-pytorch==0.3.3
+Requires-Dist: segmentation-models-pytorch==0.4.0
 Requires-Dist: ACSConv==0.1.1
 Requires-Dist: requests>=2.32.2
 Requires-Dist: docker
 Requires-Dist: dicom-anonymizer==1.0.12
 Requires-Dist: twine
-Requires-Dist: zarr
 Requires-Dist: keyring
-Requires-Dist: monai==1.3.0
+Requires-Dist: monai==1.4.0
 Requires-Dist: click>=8.0.0
 Requires-Dist: deprecated
 Requires-Dist: packaging==24.0
@@ -64,6 +67,8 @@ Requires-Dist: huggingface-hub==0.25.1
 Requires-Dist: openslide-bin
 Requires-Dist: openslide-python==1.4.1
 Requires-Dist: lion-pytorch==0.2.2
+Requires-Dist: pydantic==2.10.6
+Requires-Dist: panoptica>=1.3.3
 Dynamic: author
 Dynamic: author-email
 Dynamic: classifier
@@ -71,6 +76,7 @@ Dynamic: description
 Dynamic: description-content-type
 Dynamic: keywords
 Dynamic: license
+Dynamic: license-file
 Dynamic: requires-dist
 Dynamic: requires-python
 Dynamic: summary

{gandlf-0.1.3.dev20250319.dist-info → gandlf-0.1.4.dev20250503.dist-info}/RECORD

@@ -1,34 +1,50 @@
 GANDLF/__init__.py,sha256=8Zq1qibZtNL7X2kaR0yNLoB3j1z9p-MotzJ_fctWAoE,112
-GANDLF/config_manager.py,sha256=fglvns0FTkMQDcuhH3dXm8dqzpXzGq3-9rJ85AtFk1o,34385
-GANDLF/inference_manager.py,sha256=ncT8HFo46IIpD_mw8i2mmOnMkhvPrWJITfiNQ1uBIco,4048
+GANDLF/config_manager.py,sha256=OzillTYjIzkBE3Xb3jExU1RifIUQqt6PC5HR3Qp-_dk,2340
+GANDLF/inference_manager.py,sha256=5tCsAxJ88uafPPHz8NTPrbwLfxCFL15uiF2nJ3HgcP0,6116
 GANDLF/logger.py,sha256=oamQ1SOTTpAnC8vQ67o211Q6_bExGg_hAuqlHGlJfAI,2951
 GANDLF/logging_config.yaml,sha256=9XxRxAKtLn5ehT1khpR8wEiJGW64sx1lylAspM5KaWk,1337
 GANDLF/parseConfig.py,sha256=jO-ybIPxLw23OWDvFdTukbft0ZM8UOofGnoL2C5CEps,754
-GANDLF/training_manager.py,sha256=t7zcOVdgu5AhOhY544joxfTNR00ixnXQfQ4GWLk2o6Y,8058
-GANDLF/version.py,sha256=Bb4h93p7WC0T4d727du_nd8jFuIOo6zdW1NBqCmjNbE,34
+GANDLF/training_manager.py,sha256=AZlf-fl7KAwZgyre8-0M5lAyma6NvtiGX6XT51AJdxU,11436
+GANDLF/version.py,sha256=ZAXGLHkALqlnbxjHW0NwNNV_mWbVfN8L8YylfTOAhmc,34
 GANDLF/anonymize/__init__.py,sha256=Nxig-jM-a-aKlK09PNi1zhNulEpLTyjnsY_oGQKdjhQ,1953
 GANDLF/anonymize/convert_to_nifti.py,sha256=MOfSDncFGJGb-EQP9sFGn0yuKpX010Ioi2KNwttaex8,1339
 GANDLF/cli/__init__.py,sha256=F05eyL2HKyWkHczRZuPE_Z2Yg685P9ARYxTwz6njGeQ,784
 GANDLF/cli/config_generator.py,sha256=_j0aRV3puiIItJ1WOHRNVII_TkGJuH9ep6SuwPblcnc,4502
 GANDLF/cli/data_split_saver.py,sha256=72ygy9s3INt2NFAfSW4j9dxxmuuvfzfBTF0Hwh0nYBU,1720
-GANDLF/cli/deploy.py,sha256=55mnId66_-KOWROwPj9OEePENWm58BYFBvn_kWOdil8,15153
-GANDLF/cli/generate_metrics.py,sha256=t9Ecr8Vva03JBX_Og70QRnvJtmSXQg1q8YxlyoC_9DY,18981
+GANDLF/cli/deploy.py,sha256=7yPqRox6e5MB-VzHxmA4kt-W5dWwElxDm1mm019jN74,15155
+GANDLF/cli/generate_metrics.py,sha256=TxKJsDHtGBgKc-Fyx4F_tc_CMu5hSsO8xlc5Te86DGk,19973
 GANDLF/cli/huggingface_hub_handler.py,sha256=vGQYwDiT3Y8yk_xLNkB8DgW1rYpEcwYp0TCdCO5oWTs,4631
-GANDLF/cli/main_run.py,sha256=Du14psJsqS2wUe8XkfmdlQQutlNO7RYqoK-VNOslA0I,4488
+GANDLF/cli/main_run.py,sha256=gYNJQGvmwqBjj3ARRxj-y3R1PMOP4qTwcwFSRBbkh0k,4352
 GANDLF/cli/patch_extraction.py,sha256=30RpBfjpwE9GLRSSD7JRqwoHngESQL0eOt09n_5r18Q,3555
 GANDLF/cli/post_training_model_optimization.py,sha256=cuPqUGrFcTIfhT67lAxpnywt8t9yrlqyHUl6wjr18qo,2111
 GANDLF/cli/preprocess_and_save.py,sha256=pvOI5oaM-GLDF2TdK9F9j_KMyvypNh155Hut_AvZphg,8814
 GANDLF/cli/recover_config.py,sha256=mQb9WkyyoGArkYluTCmnM9vTBeDaRSuT2I1AI4REwQQ,1828
-GANDLF/compute/__init__.py,sha256=xOpuMrg2jvQ8DDXsdkhS1WcNpCj9XIYyeMt1X89JlUI,128
-GANDLF/compute/forward_pass.py,sha256=2W_a9rCfy5TkxUXwerSxqmwARfqZuUQcCyjX6pHMeWw,22465
-GANDLF/compute/generic.py,sha256=bqhRC5SQLgIMo62YYxBrDTkMGaZ6pRPR7mjNV5J2T-4,4295
-GANDLF/compute/inference_loop.py,sha256=JFh-7c7kiPwO3Qey5973co-wWm2n3gYNhpuDXMzv1DY,15365
-GANDLF/compute/loss_and_metric.py,sha256=dWGVPUdl76DBQsaFnrmfxYalBmwsOoOCsgJQGIYQnX8,6689
+GANDLF/compute/__init__.py,sha256=y9-astlukrCNOXcCDz68opEvyuNCjtGqB2P4yybi3MY,44
+GANDLF/compute/forward_pass.py,sha256=npSdxKSXKgoqdutFnVF3X0cBo8xnIVq5FHpLHUqw-4Q,22464
+GANDLF/compute/generic.py,sha256=vM18DoZV0cBjX8NMP-MpRqd9ufcB77pWkRDHAHCTLTI,8022
+GANDLF/compute/inference_loop.py,sha256=7yhGzxMztVxqjHIeI_lNky1CL76_4iSMdXF6BRvYFKE,15365
+GANDLF/compute/loss_and_metric.py,sha256=vKta--M_2cvsqDfeGLFcE5UZNyfHYkOyjqxlc3xZWK8,6680
 GANDLF/compute/step.py,sha256=di8zk8EgRKu5y0-Aq0aBTgrKo30SWM8wdLsnVeDaZqU,4765
-GANDLF/compute/training_loop.py,sha256=O31kQtQ3LMoOnh6yq1jiV68Yb7yDBMd6B9vvkK8gNUA,24835
+GANDLF/compute/training_loop.py,sha256=TI0aDov1K2OjbmB1LxDVtuZzbB99_wIiTu22tsH4GS0,24871
+GANDLF/configuration/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+GANDLF/configuration/default_config.py,sha256=utmsmxTzyNSOz_E7SNQCxXMjbQ3H9RII-e9iBFALrb8,3069
+GANDLF/configuration/differential_privacy_config.py,sha256=nvLoBt11VTD_pGsQD8cmmgv-XMVZR5s0I-dkKrLhRXE,584
+GANDLF/configuration/exclude_parameters.py,sha256=rg754BoaWBMmdKVgrmFSmtsw9uw7V9Yi-VXcXPp1edY,46
+GANDLF/configuration/model_config.py,sha256=CFlTU7ypX5FzPDj-UMDDQXThHkgLeqvPEtPy79dZwdU,3313
+GANDLF/configuration/nested_training_config.py,sha256=-8jygtu2ku2amSPXZ95AZTYPoqmw1O3wWH2apcsceFg,974
+GANDLF/configuration/optimizer_config.py,sha256=58MApzpHQLyTvVqNM8wHRruM4PxNK8gRfvVuUEDbcGw,3325
+GANDLF/configuration/parameters_config.py,sha256=p31KVVwWVjOSifWoXPuUKynzyeeUoTSTVdkmGHGgqpk,284
+GANDLF/configuration/patch_sampler_config.py,sha256=1CFjhB7Idfdw0ckXlyVwK-dDBKYlcuXrDi-uRdJjPA4,357
+GANDLF/configuration/post_processing_config.py,sha256=Wx1YYuKnuxP0bgN5j9rQWUk3cD4-XJ_hE3cB-7xvTpg,349
+GANDLF/configuration/pre_processing_config.py,sha256=qjaSzdsDn3M5W5j4eUN21GknTTPSFlR8b9_69dbl-m8,3322
+GANDLF/configuration/scheduler_config.py,sha256=OKhXBrDKAM1YFROnrDy7H_V5Zs2GRkUWzaVkWTVfEu0,3308
+GANDLF/configuration/user_defined_config.py,sha256=_-OejzxteociR4NM7xrtOKfWBfUmjRvNr_SMmdcZCYc,5179
+GANDLF/configuration/utils.py,sha256=p0bzaOKRvGi-beB99dBfkfl4rTDJ1F9xar78YJbii4U,3247
+GANDLF/configuration/validators.py,sha256=Mz0SmVQhF2cTyUs5Otxn_7T6L0MoaPRHPoOYimks96o,19875
 GANDLF/data/ImagesFromDataFrame.py,sha256=d4zwJbvBGxiReVw_WsWsV1v0cWKw3qKRrGXZqygOH10,13106
-GANDLF/data/__init__.py,sha256=m-jDpE3EFxj0K4eJN7WUrJMoZkP0htIfoAOeAHXB-q4,2412
+GANDLF/data/__init__.py,sha256=-ikaFkwMzI67_e1k6e-zMiF2bjGb-y8hVXeNlPWKreI,2292
 GANDLF/data/inference_dataloader_histopath.py,sha256=dOSuOnEFOcA2nPeRuUunSHIz65ea7w-iv3VTdp8E4sE,6009
+GANDLF/data/lightning_datamodule.py,sha256=Nd3NDh4T57lpW5up5ax1WHL8MYot6BCptIt_mmWshBQ,4443
 GANDLF/data/augmentation/__init__.py,sha256=fylJfhi-trBMstGTcvkwYDT6zBVXrkaobNI21ItaiGU,2255
 GANDLF/data/augmentation/blur_enhanced.py,sha256=JB-vmkE8xplx2pWOAAw1zXf7Ys8gWhC1tep0QfteyWI,2576
 GANDLF/data/augmentation/hed_augs.py,sha256=BQihsQPIlWw4LeS29RLJp0w2boCzvrw-zc3wkADKWoc,12443
@@ -71,7 +87,7 @@ GANDLF/entrypoints/optimize_model.py,sha256=w3A1n2jvH5s2bp9m3E3rUX1wd1J4Tt_zfOD3
 GANDLF/entrypoints/patch_miner.py,sha256=StrdmbAiECPpsN1AoZo75Mc1y-UAmKFyNfFmJrnHOg4,3095
 GANDLF/entrypoints/preprocess.py,sha256=-2XbLmRkQossOpZCU0HvvWWuKbefq6OE_Bf2HeHoYNk,5571
 GANDLF/entrypoints/recover_config.py,sha256=3eyYbQJMPcgZz6NDZFpJadbgUdnMF6chudRKqmBaJ4E,3712
-GANDLF/entrypoints/run.py,sha256=6A3OQw7DEeuJ_tPYnz4QovnI2sem09taAsq82gPZ6b8,9570
+GANDLF/entrypoints/run.py,sha256=V9z9Ad7EV_nehK6z9wpnic1TDMMv0KPmnmw0VPIHGJQ,8914
 GANDLF/entrypoints/split_csv.py,sha256=TZMpYjzPXD8koOdB7kokJYStQcfMzQD0z_Nysxvuxwo,3662
 GANDLF/entrypoints/subcommands.py,sha256=LyYgN5bOAF5K6laGWChe_CBi02RmPhIS7nYpiojOC0w,1768
 GANDLF/entrypoints/verify_install.py,sha256=E204smRPuH7Fv4rBPkjVXmxMPB_Tzd2lTAdVxp2hNhY,1370
@@ -79,22 +95,26 @@ GANDLF/grad_clipping/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hS
 GANDLF/grad_clipping/adaptive_gradient_clipping.py,sha256=ABt2sBLVFkW_5ffA-GkrFaxCFi2AkEOT3WjroUif2fI,2896
 GANDLF/grad_clipping/clip_gradients.py,sha256=bXXyy0suMmc0P3bj7odHVYX18w--IZ6Mu1cseFQ9F7U,1157
 GANDLF/grad_clipping/grad_scaler.py,sha256=Uwc37zkGs_513U-QOYRVNc1VUOW6rd1aGYsaXmoXPaw,3013
-GANDLF/losses/__init__.py,sha256=tt1xh66FhuYqLCx8qvjwoVZJkyUhQyqRiWDVQxhS7so,887
+GANDLF/losses/__init__.py,sha256=CEZPbqjjj3w9eCPhifmtgU1JjsVR5xnT1HCbF0cVLqo,1533
 GANDLF/losses/hybrid.py,sha256=mYR3oYsz9I3GcVKsrY7uiafAvyunfCx3rlgZPk-M9zo,1599
 GANDLF/losses/hybrid_new.py,sha256=aC_2mGrDGPQKw9pgtvwhIvsT8a9-0Y8mQo7KSBP7XbE,795
+GANDLF/losses/loss_calculators.py,sha256=U_uDyVRhBeZ0wPjDNTR39zLBE0o1hWaSU3jLSmrCmzo,2733
 GANDLF/losses/loss_interface.py,sha256=3iI9DVP5M1zrYMGEqfS3vDx46-A5Uuf73H123P2LeNo,5497
 GANDLF/losses/regression.py,sha256=iB40tuexbQS4O2WzgBtTEYYmgLpARhC_yCMi8K-IQn8,8711
 GANDLF/losses/regression_new.py,sha256=WMfJkoRRAHa2Jpjr8Zex-pHMqiiszNUnnq9H4XZQYFs,1992
-GANDLF/losses/segmentation.py,sha256=jZauVnYDcavt1PIuxe6UZYfRSdGZe1nlfz3oHbGaJ-w,12248
+GANDLF/losses/segmentation.py,sha256=txyVEgCl9mDWXxBqlHYhZI7nAwzDAelQj0MA__HBJN4,12341
 GANDLF/losses/segmentation_new.py,sha256=GK1F-5x2aaVr8zQFNUIUw5rlGydXhoXKWitiEcSctAY,6936
-GANDLF/metrics/__init__.py,sha256=4CK3J0nHk_2zrhkKFki1CIJOeEHOB8yfpp7SdzJY2Kg,3549
+GANDLF/metrics/__init__.py,sha256=k8xcpujpUg22R33keOUNetzrLnGSHsEAmClkFeDHZxc,4470
 GANDLF/metrics/classification.py,sha256=6rVZjXxHwfJFqw19UiRMYK3uF656QgfMwWDz-LOHUaw,4692
-GANDLF/metrics/generic.py,sha256=VnHRvMgKM4M_aOuZ3oYugnK7v_vkaJkH5KS4ns2eRuU,4984
+GANDLF/metrics/generic.py,sha256=N-h9sDwuAsskK7AdZ2N0otcbFTmKxh9kdyi6MQkDO60,4983
+GANDLF/metrics/metric_calculators.py,sha256=c-NAl84yniTWjMKePBlf0TsCIJnYHODplt8HeEk9Uac,3610
+GANDLF/metrics/panoptica_config_brats.yaml,sha256=W2EBe-NRHNYc7vRZdxnAyKy0SgXQ21F3K3Nzb0EvwE0,2368
 GANDLF/metrics/regression.py,sha256=Ca58jo9OL1GdjB2ZvQCcrKWz9F67iY5rcbnQxT-LpcU,4584
 GANDLF/metrics/segmentation.py,sha256=zqWEI6LuphFM0Nsm0sG2XeqAxpNJnoK1Z9q4FJt58c8,25645
+GANDLF/metrics/segmentation_panoptica.py,sha256=QpdYdHMsZhwVo9hOqB2FFEXItEu2dEn9cPw6vB5cZwU,1025
 GANDLF/metrics/synthesis.py,sha256=IfYit-R0O5ZUfmsWfLD6BqcLjjoEiotx0lUdsXWlbao,6730
 GANDLF/models/MSDNet.py,sha256=mzBkw_kQigSDTxaR7tojhdI4ECIQ65i_qiCrNaZmBHI,3442
-GANDLF/models/__init__.py,sha256=PJIbBtICuERaXCbK5u7BvEGFTje44FaDkCLpHLltthw,4063
+GANDLF/models/__init__.py,sha256=3SwskWonRW1AduPe6CYQWucU8ZBkkfJUpEF34I0NWP0,4286
 GANDLF/models/brain_age.py,sha256=itV27Bzh7a9er3WsenntvlnzCM0ZLBw1Xents4-wuAc,2084
 GANDLF/models/deep_unet.py,sha256=VyzwuEcjunO5dwU28NYyZ3XlqXszy18fGT8Kd6J_2bQ,9722
 GANDLF/models/densenet.py,sha256=9o-vJaKNfxtdhc8u71TfKelwL1d_2pn702mqj19lI5o,11848
@@ -105,6 +125,7 @@ GANDLF/models/imagenet_unet.py,sha256=vTAuckJ0Bs9nfe2G0nKISqgxWArIYdM4poGpTk56QY
 GANDLF/models/imagenet_vgg.py,sha256=FrMNLIxCtaMHWZ9JO7pWtoXTT1f8eVDuREx9108Aw2g,10367
 GANDLF/models/light_unet.py,sha256=jzcTHH7kv_ikriYxIe7fbQJkmwOQZzePW4KLCW2FGWM,6645
 GANDLF/models/light_unet_multilayer.py,sha256=xrKy--lljbCX7U2eYjpFZA69C6qlHmA5W-fEckPBll4,4534
+GANDLF/models/lightning_module.py,sha256=kZhlCSD_JLJFFqVRGUEz1CgvENWHqVgc418_fvbQdkk,83347
 GANDLF/models/modelBase.py,sha256=NFmKcEylNchm0LVqwRoD_iIMFj36IsaKwVCPJPLux3g,5752
 GANDLF/models/resnet.py,sha256=0PB7sGSLHUq2F6tyHYqxepmU1eBccSwCzBrD7Qmc280,16987
 GANDLF/models/sdnet.py,sha256=uhHvOHRTAx4ANxv2lKXfBn6Bg4Broqzb77qekVjXeFI,14528
@@ -132,7 +153,7 @@ GANDLF/models/seg_modules/add_conv_block.py,sha256=ww8pWBq6r_DYwe8DHLZM2oPXmJOIU
 GANDLF/models/seg_modules/add_downsample_conv_block.py,sha256=UUvMjFZSvfwSxEqwS1uonhPaFHc1Y39aKx56tWWrrqk,410
 GANDLF/models/seg_modules/average_pool.py,sha256=8TRsq5ScmofJXumS-MEWZ0We0EpYVX44rzKqLad9uEg,1980
 GANDLF/models/seg_modules/out_conv.py,sha256=mOfHlIaznx8pauMkAiu_Ni57383YYJtm5E1ut7--ks0,3379
-GANDLF/optimizers/__init__.py,sha256=dri2oPKH6VYSqJyEgkjBQP9_jFibbWjmzZwJgHwF8kg,1433
+GANDLF/optimizers/__init__.py,sha256=cJiIvV0zT-b0grrXnvFm8NOOXgX2cPgGoyc8fdn9cYE,1281
 GANDLF/optimizers/wrap_monai.py,sha256=bDB055PZYj44wyBNXU3oYxjBPa4fyv_5BsRpPtkIUf0,782
 GANDLF/optimizers/wrap_torch.py,sha256=phDMJ-BPSH8nsfqcJLUndhQEb021Uf1CicpAB5C-gvY,8785
 GANDLF/optimizers/thirdparty/__init__.py,sha256=o9R9I2dpsd0pPKf7x-idATt5eadhoIX00Zj6p8IUhYs,105
@@ -143,25 +164,27 @@ GANDLF/privacy/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 GANDLF/privacy/opacus/__init__.py,sha256=8FkYz9ViiEUIF9JF3t5rbJjjOVUnNomnYR-bd500kZI,173
 GANDLF/privacy/opacus/config_parsing.py,sha256=1x9XTrRA-zVJI8dieVDECL-RWd4SlOsk639gNObZMnI,2367
 GANDLF/privacy/opacus/model_handling.py,sha256=LSFo_TtCHwFhThJ9i7agWFjxVx1aABHC3bIUVLReFPs,6437
+GANDLF/privacy/opacus/opacus_anonymization_manager.py,sha256=8VzMtezagKlkaI6x-UL7NPiXIIFQ014Iw9atsCy_46w,9549
 GANDLF/privacy/opacus/training_utils.py,sha256=ejXUIhLNjRzTebF-7uCYUluK7jyINDOpLe2yhOKacJg,3683
-GANDLF/schedulers/__init__.py,sha256=y70gA0v8DTNYYCEr8zbc3WB-x2xkZ7VzUvdSznTCTMI,1123
+GANDLF/schedulers/__init__.py,sha256=jMTea7kX7zawFrtjVDKHDLIfBgt7GD-G2z9_eKjI1_A,1320
 GANDLF/schedulers/wrap_monai.py,sha256=_kP3ArUIFK3x3u0nQV8Gi2oPruMIGGTSZxa4bSygGPs,400
 GANDLF/schedulers/wrap_torch.py,sha256=NuCAmYMKgYeFzO1PEiHfnFaHpzuQU0DsK5NOi-gVxOo,6746
-GANDLF/utils/__init__.py,sha256=yAXxhPgSu0FdUcvocaW2cGSljBwrIjRMjALpJcMWjbY,1677
+GANDLF/utils/__init__.py,sha256=QpNBl-fTrFUDyChAnaidPXdT-pY080pMG9kUTrlJ9OY,1691
 GANDLF/utils/data_splitter.py,sha256=IBVsLYo4y28G6LedbLPYWyiAUiaLxD4wWZD4EZC2D6I,10174
 GANDLF/utils/exceptions.py,sha256=SmEGzCbc5mXxjylmA9XE4PXZdmkkfOjrEh473cWTAvA,100
 GANDLF/utils/gandlf_logging.py,sha256=a1yV4244BGO5gp4OuAIKJUZgR21U1OPdovJ5nH1uUrs,2162
 GANDLF/utils/generic.py,sha256=NyJgaeCdryBlHMziPPLmWtIk450u042NWQgs_qh7knw,11744
 GANDLF/utils/handle_collisions.py,sha256=UFgfrddQqeipOdpKk0Mo-wwJ-TYTESNZyXzEyQeaLdI,3082
-GANDLF/utils/imaging.py,sha256=qetUoX7hfgvwtIkSkOIgaCdcHZIkuT5NR2A0fnJ_PHc,10354
+GANDLF/utils/imaging.py,sha256=FZYtRXMHYUkL_7bzVtwsnrYyrYAm6Hvs5R-tFnEbYkY,15083
 GANDLF/utils/modelbase.py,sha256=MJ1ufJGQ-ZgfYWTg7o4f-Iz78d7SYiyxN7MT8bnNDaw,950
-GANDLF/utils/modelio.py,sha256=tWeTwVTvQy5I5hydxNZz4N8xEU3lHMayMa9gx5-Vr7Q,8016
+GANDLF/utils/modelio.py,sha256=9koLUn3wCODtNOYA2abtxkqxfLgEG07Bhh-NYIrtWOc,8084
 GANDLF/utils/parameter_processing.py,sha256=DA7ZEsizWWLJZnCxBnDNh1NyA-bw5oirOvodiZZWbIk,5675
+GANDLF/utils/pred_target_processors.py,sha256=aatXJ6jdJaNAHa_tPzHfC1gOQrPYJLtg-cYeUFvkM_s,2701
 GANDLF/utils/tensor.py,sha256=AOwNTQfw9mnsomGwOF2Q_rdDS1WANlIatB0hhZN0gSg,21504
-GANDLF/utils/write_parse.py,sha256=bTBNxIaIGgjusq1mlHA8hioKZ6XmEJWFhfZWAQz2a9I,9197
-gandlf-0.1.3.dev20250319.dist-info/LICENSE,sha256=GlZPAfA4eckod8IVayhBXkqCpESXf6cc9BGli_Jwims,11357
-gandlf-0.1.3.dev20250319.dist-info/METADATA,sha256=jre8HnA6_GaaUkvo5WtgmBikrNfI36H616f7-kp3z0c,9726
-gandlf-0.1.3.dev20250319.dist-info/WHEEL,sha256=beeZ86-EfXScwlR_HKu4SllMC9wUEj_8Z_4FJ3egI2w,91
-gandlf-0.1.3.dev20250319.dist-info/entry_points.txt,sha256=agwocNI7Upi-sFDe1rMl71dGN8YhCBB7WJmtBHRF4jg,902
-gandlf-0.1.3.dev20250319.dist-info/top_level.txt,sha256=i5D9wEbQhl085_9Lx2m7x-9Zu6nlx1tjYYbuSihG09E,7
-gandlf-0.1.3.dev20250319.dist-info/RECORD,,
+GANDLF/utils/write_parse.py,sha256=HROxskhet-uIdUJmz16z_7p9r0mf8hWsqQFygwZ8ap0,9202
+gandlf-0.1.4.dev20250503.dist-info/licenses/LICENSE,sha256=GlZPAfA4eckod8IVayhBXkqCpESXf6cc9BGli_Jwims,11357
+gandlf-0.1.4.dev20250503.dist-info/METADATA,sha256=y7Nz_2o6DcvEYOsUlI-3rMOx26iJZvzLq8owya8Ltsc,9932
+gandlf-0.1.4.dev20250503.dist-info/WHEEL,sha256=7ciDxtlje1X8OhobNuGgi1t-ACdFSelPnSmDPrtlobY,91
+gandlf-0.1.4.dev20250503.dist-info/entry_points.txt,sha256=agwocNI7Upi-sFDe1rMl71dGN8YhCBB7WJmtBHRF4jg,902
+gandlf-0.1.4.dev20250503.dist-info/top_level.txt,sha256=i5D9wEbQhl085_9Lx2m7x-9Zu6nlx1tjYYbuSihG09E,7
+gandlf-0.1.4.dev20250503.dist-info/RECORD,,

{gandlf-0.1.3.dev20250319.dist-info → gandlf-0.1.4.dev20250503.dist-info}/WHEEL

@@ -1,5 +1,5 @@
 Wheel-Version: 1.0
-Generator: setuptools (76.1.0)
+Generator: setuptools (80.2.0)
 Root-Is-Purelib: true
 Tag: py3-none-any