cache examples false

app.py

@@ -1,3 +1,4 @@
+import unittest
 import gradio as gr
 import torch
 import cv2
@@ -6,6 +7,11 @@ from albumentations.pytorch import ToTensorV2
 import matplotlib.pyplot as plt
 import matplotlib
 matplotlib.use('agg')
+import torch    
+import cv2
+from pytorch_grad_cam import EigenCAM
+from pytorch_grad_cam.utils.image import show_cam_on_image
+
 
 import config
 from model import YOLOv3
@@ -15,37 +21,30 @@ from utils import (
     plot_image
 )
 
-# fname = 'epoch=36-step=19166.ckpt'
-fname = 'yolov3.pth'
+def yolov3_reshape_transform(tensor, ):
+    return tensor[0]
+
+# fname = 'epoch=38-step=20202.ckpt'
 # checkpoint = torch.load(fname, map_location=torch.device('cpu'))
 # model_state_dict = checkpoint['state_dict']
-# torch.save(model.state_dict(), 'yolov3.pth')
 model = YOLOv3(num_classes=20)
+# model.load_state_dict(model_state_dict)
+# torch.save(model.state_dict(), 'yolov3.pth')
+fname = 'yolov3.pth'
 model.load_state_dict(torch.load(fname))
 
-IMAGE_SIZE = 416
+IMAGE_SIZE = config.IMAGE_SIZE
 S = [IMAGE_SIZE // 32, IMAGE_SIZE // 16, IMAGE_SIZE // 8]
 anchors = ( torch.tensor(config.ANCHORS)
     * torch.tensor(config.S).unsqueeze(1)\
         .unsqueeze(1).repeat(1, 3, 2)
     )
 
-test_transforms = A.Compose(
-    [
-        A.LongestMaxSize(max_size=IMAGE_SIZE),
-        A.PadIfNeeded(
-            min_height=IMAGE_SIZE, min_width=IMAGE_SIZE, border_mode=cv2.BORDER_CONSTANT
-        ),
-        A.Normalize(mean=[0, 0, 0], std=[1, 1, 1], max_pixel_value=255,),
-        ToTensorV2(),
-    ],
-)
-def object_detector(input_image):
-    input_img = test_transforms(image=input_image)['image']
+def object_detector(input_image, thresh = 0.8, iou_thresh = 0.5):
+    input_img = config.test_transforms(image=input_image)['image']
     input_img = input_img.unsqueeze(0)
 
-    thresh = 0.6
-    iou_thresh = 0.5
+
     with torch.no_grad():
         out = model(input_img)
         bboxes = []
@@ -54,9 +53,9 @@ def object_detector(input_image):
             anchor = anchors[i]
             bboxes += cells_to_bboxes(
                 out[i], anchor, S=S, is_preds=True
-            )
+            )[0]
         nms_boxes = non_max_suppression(
-            bboxes[0], iou_threshold=iou_thresh, 
+            bboxes, iou_threshold=iou_thresh, 
             threshold=thresh, box_format="midpoint",
         )
         fig = plot_image(input_img.squeeze(0).permute(1,2,0).detach().cpu(),
@@ -67,27 +66,75 @@ def object_detector(input_image):
         image_path = "plot.png"
         fig.savefig(image_path)
         plt.close()
-        return gr.update(value=image_path, visible=True)
+
+        # target_layers = [model.layers[21]]
+        # cam = EigenCAM(model, target_layers, use_cuda=False,
+        #                reshape_transform=yolov3_reshape_transform,
+        #                )
+        # grayscale_cam = cam(input_img, target_layers)[0][0, :, :]
+        # cam_image = show_cam_on_image(img, grayscale_cam, use_rgb=True)
+        return gr.update(value=image_path, visible=True),\
+            gr.update(value=image_path, visible=True)
 
 # Define the input and output components for Gradio
 input_image = gr.Image(label="Input image")
-output_box = gr.Image(label="Output image")\
+confidence_level = gr.Slider(0.5, 1, value=0.6, step=0.01, 
+                    label="confidence level")
+iou_level = gr.Slider(0.5, 1, value=0.6, step=0.01, 
+                    label="Interference over union level")
+output_box = gr.Image(label="Output image", visible=False,)\
+                   .style(width=428, height=428)
+cam_output = gr.Image(label="cam output", visible=False)\
                    .style(width=428, height=428)
 images_path = "examples/"
 
+gr_interface = gr.Interface(
+    fn=object_detector,
+    inputs=[input_image, confidence_level, iou_level],
+    outputs=[output_box, cam_output],
+    examples=[[images_path + "000015.jpg"], 
+               [images_path + "000017.jpg"],
+               [images_path + "000030.jpg"],
+               [images_path + "000069.jpg"],
+               [images_path + "000071.jpg"],
+               [images_path + "000084.jpg"],
+               [images_path + "000086.jpg"],
+               [images_path + "000088.jpg"],
+               [images_path + "000100.jpg"],
+               ],
+    cache_examples=False
+                 )
 
-# Create the Gradio interface
-gr.Interface(fn=object_detector, inputs=input_image, outputs=output_box,
-             examples=[[images_path + "000015.jpg"], 
-                       [images_path + "000017.jpg"],
-                       [images_path + "000030.jpg"],
-                       [images_path + "000069.jpg"],
-                       [images_path + "000071.jpg"],
-                       [images_path + "000084.jpg"],
-                       [images_path + "000086.jpg"],
-                       [images_path + "000088.jpg"],
-                       [images_path + "000095.jpg"],
-                       [images_path + "000100.jpg"],
-                       ],
-             ).launch()
+gr_interface.launch()
+
+# class TestGradioInterfaceInput(unittest.TestCase):
+#     def test_valid_image_input(self):
+#         # Create a valid input image
+#         input_image = images_path + "000015.jpg"
+#
+#         # Pass the image through the interface
+#         output = gr_interface(input_image)
+#
+#         # Assert the output matches the expected result
+#         self.assertEqual(output[0].shape, (3, 416, 416))
+#
+# if __name__ == '__main__':
+#     unittest.main()
+#
 
+# Create the Gradio interface
+# gr.Interface(fn=object_detector, inputs=input_image, 
+#              outputs=[output_box, cam_output],
+#              examples=[[images_path + "000015.jpg"], 
+#                        [images_path + "000017.jpg"],
+#                        [images_path + "000030.jpg"],
+#                        [images_path + "000069.jpg"],
+#                        [images_path + "000071.jpg"],
+#                        [images_path + "000084.jpg"],
+#                        [images_path + "000086.jpg"],
+#                        [images_path + "000088.jpg"],
+#                        [images_path + "000095.jpg"],
+#                        [images_path + "000100.jpg"],
+#                        ],
+#              ).launch()
+#

config.py

@@ -26,37 +26,6 @@ ANCHORS = [
 means = [0.485, 0.456, 0.406]
 
 scale = 1.1
-train_transforms = A.Compose(
-    [
-        A.LongestMaxSize(max_size=int(IMAGE_SIZE * scale)),
-        A.PadIfNeeded(
-            min_height=int(IMAGE_SIZE * scale),
-            min_width=int(IMAGE_SIZE * scale),
-            border_mode=cv2.BORDER_CONSTANT,
-        ),
-        A.Rotate(limit = 10, interpolation=1, border_mode=4),
-        A.RandomCrop(width=IMAGE_SIZE, height=IMAGE_SIZE),
-        A.ColorJitter(brightness=0.6, contrast=0.6, saturation=0.6, hue=0.6, p=0.4),
-        A.OneOf(
-            [
-                A.ShiftScaleRotate(
-                    rotate_limit=20, p=0.5, border_mode=cv2.BORDER_CONSTANT
-                ),
-                # A.Affine(shear=15, p=0.5, mode="constant"),
-            ],
-            p=1.0,
-        ),
-        A.HorizontalFlip(p=0.5),
-        A.Blur(p=0.1),
-        A.CLAHE(p=0.1),
-        A.Posterize(p=0.1),
-        A.ToGray(p=0.1),
-        A.ChannelShuffle(p=0.05),
-        A.Normalize(mean=[0, 0, 0], std=[1, 1, 1], max_pixel_value=255,),
-        ToTensorV2(),
-    ],
-    bbox_params=A.BboxParams(format="yolo", min_visibility=0.4, label_fields=[],),
-)
 test_transforms = A.Compose(
     [
         A.LongestMaxSize(max_size=IMAGE_SIZE),
@@ -66,7 +35,6 @@ test_transforms = A.Compose(
         A.Normalize(mean=[0, 0, 0], std=[1, 1, 1], max_pixel_value=255,),
         ToTensorV2(),
     ],
-    bbox_params=A.BboxParams(format="yolo", min_visibility=0.4, label_fields=[]),
 )
 
 PASCAL_CLASSES = [

requirement.txt

@@ -1,6 +0,0 @@
-torch
-torchvision
-grad-cam
-pillow
-numpy
-albumentations

yolov3.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:af069adc24fec136b92b068e6e6c1361dd4b9dba7797dac7798489e8181d021c
+oid sha256:8413ddf1b2ee36957265778c668a7592fb92704b5ba3692ab2e3137e89571649
 size 246865311