Update app.py

app.py

@@ -2,50 +2,24 @@ from huggingface_hub import from_pretrained_fastai
 import gradio as gr
 from fastai.vision.all import *
 from icevision.all import *
+from icevision.models.checkpoint import *
 import PIL
 
-class_map = ['raccoon']
+checkpoint_path = "fasterRCNN_resnet18_Raccoons.pth"
+model = models.torchvision.faster_rcnn.model(backbone=models.torchvision.faster_rcnn.backbones.resnet18_fpn(pretrained=True),
+                                       num_classes=1)
 
-presize = 512
-size = 384
+state_dict = torch.load(checkpoint_path, map_location=torch.device('cpu'))
+model.load_state_dict(state_dict)
 
-train_tfms = tfms.A.Adapter(
-    [*tfms.A.aug_tfms(size=size, presize=presize), tfms.A.Normalize()]
-)
-valid_tfms = tfms.A.Adapter([*tfms.A.resize_and_pad(size), tfms.A.Normalize()])
+infer_tfms = tfms.A.Adapter([*tfms.A.resize_and_pad(384),tfms.A.Normalize()])
 
-model1 = models.torchvision.faster_rcnn.model(backbone=models.torchvision.faster_rcnn.backbones.resnet18_fpn(pretrained=False), num_classes=2)
-state_dict = torch.load('fasterRCNN_resnet18_Raccoons.pth', map_location=torch.device('cpu'))
-model1.load_state_dict(state_dict)
-
-def show_preds(input_image, display_label, display_bbox, detection_threshold):
-
-    if detection_threshold==0: detection_threshold=0.5
-
-    img = PIL.Image.fromarray(input_image, 'RGB')
-
-    pred_dict  = models.torchvision.faster_rcnn.end2end_detect(img, valid_tfms, model1, class_map=class_map, detection_threshold=detection_threshold,
-                                           display_label=display_label, display_bbox=display_bbox, return_img=True, 
-                                           font_size=16, label_color="#FF59D6")
-
-    return pred_dict['img']
-
-# display_chkbox = gr.inputs.CheckboxGroup(["Label", "BBox"], label="Display", default=True)
-display_chkbox_label = gr.inputs.Checkbox(label="Label", default=True)
-display_chkbox_box = gr.inputs.Checkbox(label="Box", default=True)
-
-detection_threshold_slider = gr.inputs.Slider(minimum=0, maximum=1, step=0.1, default=0.5, label="Detection Threshold")
-
-outputs = gr.outputs.Image(type="pil")
-
-# Option 1: Get an image from local drive
-gr_interface = gr.Interface(fn=show_preds, inputs=["image", display_chkbox_label, display_chkbox_box,  detection_threshold_slider], outputs=outputs, examples=['raccoon1.jpg','raccoon2.jpg'])
-
-# #  Option 2: Grab an image from a webcam
-# gr_interface = gr.Interface(fn=show_preds, inputs=["webcam", display_chkbox_label, display_chkbox_box,  detection_threshold_slider], outputs=outputs, title='IceApp - COCO', live=False)
-
-# #  Option 3: Continuous image stream from the webcam
-# gr_interface = gr.Interface(fn=show_preds, inputs=["webcam", display_chkbox_label, display_chkbox_box,  detection_threshold_slider], outputs=outputs, title='IceApp - COCO', live=True)
-
-
-gr_interface.launch(inline=False, share=False, debug=True)
+# Definimos una función que se encarga de llevar a cabo las predicciones
+def predict(img):
+    img = PIL.Image.fromarray(img, "RGB")
+    pred_dict =  models.torchvision.faster_rcnn.end2end_detect(img, infer_tfms, model.to("cpu"), class_map=ClassMap(['raccoon']), detection_threshold=0.5)
+    return pred_dict["img"]
+    
+# Creamos la interfaz y la lanzamos. 
+gr.Interface(fn=predict, inputs=gr.inputs.Image(shape=(128, 128)), outputs=[gr.outputs.Image(type="pil", label="VFNet Inference")],
+             examples=['raccoon1.jpg','raccoon2.jpg']).launch(share=False)