Minor updates to finetuning notebook

micro_sam/training/training.py

@@ -389,7 +389,6 @@ def default_sam_dataset(
     is_train: bool = True,
     min_size: int = 25,
     max_sampling_attempts: Optional[int] = None,
-    is_seg_dataset: Optional[bool] = None,
     **kwargs,
 ) -> Dataset:
     """Create a PyTorch Dataset for training a SAM model.
@@ -413,11 +412,10 @@ def default_sam_dataset(
         is_train: Whether this dataset is used for training or validation.
         min_size: Minimal object size. Smaller objects will be filtered.
         max_sampling_attempts: Number of sampling attempts to make from a dataset.
-        is_seg_dataset: Whether the dataset is built 'from torch_em.data import SegmentationDataset'
-            or 'from torch_em.data import ImageCollectionDataset'
+        kwargs: Additional keyword arguments for `torch_em.default_segmentation_dataset`.
 
     Returns:
-        The dataset.
+        The segmentation dataset.
     """
 
     # Set the data transformations.
@@ -446,18 +444,29 @@ def default_sam_dataset(
     # Set a minimum number of samples per epoch.
     if n_samples is None:
         loader = torch_em.default_segmentation_loader(
-            raw_paths, raw_key, label_paths, label_key, batch_size=1,
-            patch_shape=patch_shape, ndim=2, is_seg_dataset=is_seg_dataset,
+            raw_paths=raw_paths,
+            raw_key=raw_key,
+            label_paths=label_paths,
+            label_key=label_key,
+            batch_size=1,
+            patch_shape=patch_shape,
+            ndim=2,
+            **kwargs
         )
         n_samples = max(len(loader), 100 if is_train else 5)
 
     dataset = torch_em.default_segmentation_dataset(
-        raw_paths, raw_key, label_paths, label_key,
+        raw_paths=raw_paths,
+        raw_key=raw_key,
+        label_paths=label_paths,
+        label_key=label_key,
         patch_shape=patch_shape,
-        raw_transform=raw_transform, label_transform=label_transform,
-        with_channels=with_channels, ndim=2,
-        sampler=sampler, n_samples=n_samples,
-        is_seg_dataset=is_seg_dataset,
+        raw_transform=raw_transform,
+        label_transform=label_transform,
+        with_channels=with_channels,
+        ndim=2,
+        sampler=sampler,
+        n_samples=n_samples,
         **kwargs,
     )
 
@@ -472,10 +481,23 @@ def default_sam_dataset(
 
 
 def default_sam_loader(**kwargs) -> DataLoader:
-    ds_kwargs, loader_kwargs = split_kwargs(default_sam_dataset, **kwargs)
+    """Create a PyTorch DataLoader for training a SAM model.
+
+    Args:
+        kwargs: Keyword arguments for `micro_sam.training.default_sam_dataset` or for the PyTorch DataLoader.
+
+    Returns:
+        The DataLoader.
+    """
+    sam_ds_kwargs, extra_kwargs = split_kwargs(default_sam_dataset, **kwargs)
+
+    # There might be additional parameters supported by `torch_em.default_segmentation_dataset`,
+    # which the users can provide to get their desired segmentation dataset.
+    extra_ds_kwargs, loader_kwargs = split_kwargs(torch_em.default_segmentation_dataset, **extra_kwargs)
+    ds_kwargs = {**sam_ds_kwargs, **extra_ds_kwargs}
+
     ds = default_sam_dataset(**ds_kwargs)
-    loader = torch_em.segmentation.get_data_loader(ds, **loader_kwargs)
-    return loader
+    return torch_em.segmentation.get_data_loader(ds, **loader_kwargs)
 
 
 CONFIGURATIONS = {
@@ -517,7 +539,7 @@ def train_sam_for_configuration(
         model_type: Over-ride the default model type.
             This can be used to use one of the micro_sam models as starting point
             instead of a default sam model.
-        kwargs: Additional keyword parameterts that will be passed to `train_sam`.
+        kwargs: Additional keyword parameters that will be passed to `train_sam`.
     """
     if configuration in CONFIGURATIONS:
         train_kwargs = CONFIGURATIONS[configuration]

workshops/i2k_2024/README.md

@@ -56,7 +56,7 @@ $ git clone https://github.com/computational-cell-analytics/micro-sam
 then go to this directory:
 
 ```bash
-$ cd micro_sam/workshops/i2k_2024
+$ cd micro-sam/workshops/i2k_2024
 ```
 
 and download the precomputed embeddings: