feat: Removes warnings for torchvision and matplotlib (#224)

* feat: Removes torchvision warnings

* feat: Removes matplotlib warnings

Author: frgfm

pyproject.toml

@@ -37,7 +37,7 @@ dependencies = [
     # cf. https://github.com/pytorch/vision/issues/4934
     # https://github.com/frgfm/Holocron/security/dependabot/5
     "Pillow>=8.4.0,!=9.2.0",
-    "matplotlib>=3.0.0,<4.0.0",
+    "matplotlib>=3.7.0,<4.0.0",
 ]
 
 [project.optional-dependencies]

tests/test_methods_activation.py

@@ -6,7 +6,7 @@
 
 
 def test_base_cam_constructor(mock_img_model):
-    model = mobilenet_v2(pretrained=False).eval()
+    model = mobilenet_v2(weights=None).eval()
     for p in model.parameters():
         p.requires_grad_(False)
     # Check that multiple target layers is disabled for base CAM
@@ -39,7 +39,7 @@ def _verify_cam(activation_map, output_size):
     ],
 )
 def test_img_cams(cam_name, target_layer, fc_layer, num_samples, output_size, batch_size, mock_img_tensor):
-    model = mobilenet_v2(pretrained=False).eval()
+    model = mobilenet_v2(weights=None).eval()
     for p in model.parameters():
         p.requires_grad_(False)
     kwargs = {}

tests/test_methods_gradient.py

@@ -26,7 +26,7 @@ def _verify_cam(activation_map, output_size):
     ],
 )
 def test_img_cams(cam_name, target_layer, output_size, batch_size, mock_img_tensor):
-    model = mobilenet_v2(pretrained=False).eval()
+    model = mobilenet_v2(weights=None).eval()
     for p in model.parameters():
         p.requires_grad_(False)
 
@@ -79,7 +79,7 @@ def test_video_cams(cam_name, target_layer, output_size, mock_video_model, mock_
 
 
 def test_smoothgradcampp_repr():
-    model = mobilenet_v2(pretrained=False).eval()
+    model = mobilenet_v2(weights=None).eval()
 
     # Hook the corresponding layer in the model
     with gradient.SmoothGradCAMpp(model, "features.18.0") as extractor:

tests/test_metrics.py

@@ -8,7 +8,7 @@
 
 
 def test_classification_metric():
-    model = mobilenet_v3_small(pretrained=False)
+    model = mobilenet_v3_small(weights=None)
     with LayerCAM(model, "features.12") as extractor:
         metric = metrics.ClassificationMetric(extractor, partial(torch.softmax, dim=-1))
 

torchcam/utils.py

@@ -4,7 +4,7 @@
 # See LICENSE or go to <https://www.apache.org/licenses/LICENSE-2.0> for full license details.
 
 import numpy as np
-from matplotlib import cm
+from matplotlib import colormaps as cm
 from PIL import Image