Fixed CUDA Execution Provider

README.rst

@@ -64,7 +64,7 @@ Now, let's dive into the details.
 They specify the path to the dataset (or MRI) and the pattern to find the files.
 
 All parameters starting with ``roiloc.`` are directly linked to our home-made ROI location algorithm.
-You can find more information about it in the `related GitHub repository <https://github.com/clementpoiret/HSF>`_.
+You can find more information about it in the `related GitHub repository <https://github.com/clementpoiret/ROILoc>`_.
 
 To date, we propose 4 different segmentation algorithms (from the fastest to the most accurate):
 

hsf/__init__.py

@@ -1 +1 @@
-__version__ = '0.1.0'
+__version__ = '0.1.1'

hsf/segment.py

@@ -61,7 +61,12 @@ def get_inference_sessions(models_path: PosixPath) -> list:
     p = Path(models_path).expanduser()
     models = list(p.glob("*.onnx"))
 
-    return [ort.InferenceSession(str(model_path)) for model_path in models]
+    ep_list = ['CUDAExecutionProvider', 'CPUExecutionProvider']
+
+    return [
+        ort.InferenceSession(str(model_path), providers=ep_list)
+        for model_path in models
+    ]
 
 
 def to_ca_mode(logits: torch.Tensor, ca_mode: str = "1/2/3") -> torch.Tensor:

pyproject.toml

@@ -1,6 +1,6 @@
 [tool.poetry]
 name = "HSF"
-version = "0.1.0"
+version = "0.1.1"
 description = "Simple yet exhaustive segmentation tool of the Hippocampal Subfields in T1w and T2w MRIs."
 authors = ["Clément POIRET <[email protected]>"]
 license = "MIT"

tests/test_hsf.py

@@ -2,4 +2,4 @@
 
 
 def test_version():
-    assert __version__ == '0.1.0'
+    assert __version__ == '0.1.1'