Revert "Monkeypatch everything"

This reverts commit e3e8d7d042.

tests/models/test_resnet.py

@@ -2,7 +2,6 @@
 # Licensed under the MIT License.
 
 from pathlib import Path
-from typing import Any, Dict
 
 import pytest
 import timm
@@ -14,11 +13,6 @@ from torchvision.models._api import WeightsEnum
 from torchgeo.models import ResNet18_Weights, ResNet50_Weights, resnet18, resnet50
 
 
-def load(url: str, *args: Any, **kwargs: Any) -> Dict[str, Any]:
-    state_dict: Dict[str, Any] = torch.load(url)
-    return state_dict
-
-
 class TestResNet18:
     @pytest.fixture(params=[*ResNet18_Weights])
     def weights(self, request: SubRequest) -> WeightsEnum:
@@ -31,8 +25,7 @@ class TestResNet18:
         path = tmp_path / f"{weights}.pth"
         model = timm.create_model("resnet18", in_chans=weights.meta["in_chans"])
         torch.save(model.state_dict(), path)
-        monkeypatch.setattr(weights, "url", str(path))
-        monkeypatch.setattr(torch.hub, "load_state_dict_from_url", load)
+        monkeypatch.setattr(weights, "url", path.as_uri())
         return weights
 
     def test_resnet(self) -> None:
@@ -58,8 +51,7 @@ class TestResNet50:
         path = tmp_path / f"{weights}.pth"
         model = timm.create_model("resnet50", in_chans=weights.meta["in_chans"])
         torch.save(model.state_dict(), path)
-        monkeypatch.setattr(weights, "url", str(path))
-        monkeypatch.setattr(torch.hub, "load_state_dict_from_url", load)
+        monkeypatch.setattr(weights, "url", path.as_uri())
         return weights
 
     def test_resnet(self) -> None:

tests/models/test_vit.py

@@ -2,7 +2,6 @@
 # Licensed under the MIT License.
 
 from pathlib import Path
-from typing import Any, Dict
 
 import pytest
 import timm
@@ -14,11 +13,6 @@ from torchvision.models._api import WeightsEnum
 from torchgeo.models import ViTSmall16_Weights, vit_small_patch16_224
 
 
-def load(url: str, *args: Any, **kwargs: Any) -> Dict[str, Any]:
-    state_dict: Dict[str, Any] = torch.load(url)
-    return state_dict
-
-
 class TestViTSmall16:
     @pytest.fixture(params=[*ViTSmall16_Weights])
     def weights(self, request: SubRequest) -> WeightsEnum:
@@ -33,8 +27,7 @@ class TestViTSmall16:
             weights.meta["model"], in_chans=weights.meta["in_chans"]
         )
         torch.save(model.state_dict(), path)
-        monkeypatch.setattr(weights, "url", str(path))
-        monkeypatch.setattr(torch.hub, "load_state_dict_from_url", load)
+        monkeypatch.setattr(weights, "url", path.as_uri())
         return weights
 
     def test_vit(self) -> None:

tests/trainers/test_byol.py

@@ -23,11 +23,6 @@ from torchgeo.trainers.byol import BYOL, SimCLRAugmentation
 from .test_utils import SegmentationTestModel
 
 
-def load(url: str, *args: Any, **kwargs: Any) -> Dict[str, Any]:
-    state_dict: Dict[str, Any] = torch.load(url)
-    return state_dict
-
-
 class TestBYOL:
     def test_custom_augment_fn(self) -> None:
         backbone = resnet18()
@@ -84,8 +79,7 @@ class TestBYOLTask:
         path = tmp_path / f"{weights}.pth"
         model = timm.create_model("resnet18", in_chans=weights.meta["in_chans"])
         torch.save(model.state_dict(), path)
-        monkeypatch.setattr(weights, "url", str(path))
-        monkeypatch.setattr(torch.hub, "load_state_dict_from_url", load)
+        monkeypatch.setattr(weights, "url", path.as_uri())
         return weights
 
     def test_weight_file(self, model_kwargs: Dict[str, Any], checkpoint: str) -> None:

tests/trainers/test_classification.py

@@ -31,11 +31,6 @@ def create_model(*args: Any, **kwargs: Any) -> Module:
     return ClassificationTestModel(**kwargs)
 
 
-def load(url: str, *args: Any, **kwargs: Any) -> Dict[str, Any]:
-    state_dict: Dict[str, Any] = torch.load(url)
-    return state_dict
-
-
 class TestClassificationTask:
     @pytest.mark.parametrize(
         "name,classname",
@@ -107,8 +102,7 @@ class TestClassificationTask:
         path = tmp_path / f"{weights}.pth"
         model = timm.create_model("resnet18", in_chans=weights.meta["in_chans"])
         torch.save(model.state_dict(), path)
-        monkeypatch.setattr(weights, "url", str(path))
-        monkeypatch.setattr(torch.hub, "load_state_dict_from_url", load)
+        monkeypatch.setattr(weights, "url", path.as_uri())
         return weights
 
     def test_weight_file(self, model_kwargs: Dict[str, Any], checkpoint: str) -> None:

tests/trainers/test_regression.py

@@ -20,11 +20,6 @@ from torchgeo.trainers import RegressionTask
 from .test_utils import RegressionTestModel
 
 
-def load(url: str, *args: Any, **kwargs: Any) -> Dict[str, Any]:
-    state_dict: Dict[str, Any] = torch.load(url)
-    return state_dict
-
-
 class TestRegressionTask:
     @pytest.mark.parametrize(
         "name,classname",
@@ -88,8 +83,7 @@ class TestRegressionTask:
         path = tmp_path / f"{weights}.pth"
         model = timm.create_model("resnet18", in_chans=weights.meta["in_chans"])
         torch.save(model.state_dict(), path)
-        monkeypatch.setattr(weights, "url", str(path))
-        monkeypatch.setattr(torch.hub, "load_state_dict_from_url", load)
+        monkeypatch.setattr(weights, "url", path.as_uri())
         return weights
 
     def test_weight_file(self, model_kwargs: Dict[str, Any], checkpoint: str) -> None: