Update app.py

app.py

@@ -1,46 +1,41 @@
 import gradio as gr
 from fastai.vision.all import *
-
-from pathlib import Path
-import PIL
 import torchvision.transforms as transforms
-
-
+import torch
+from PIL import Image
+import numpy as np
 
 device = torch.device("cuda" if torch.cuda.is_available() else "cpu") 
-model = torch.jit.load("unet.pth")
-model = model.cpu()
+model = torch.jit.load("unet.pth").to(device)
 model.eval()
 
 def transform_image(image):
-    my_transforms = transforms.Compose([transforms.ToTensor(),
-                                        transforms.Normalize(
-                                            [0.485, 0.456, 0.406],
-                                            [0.229, 0.224, 0.225])])
-    image_aux = image
+    # Definimos las transformaciones necesarias para la imagen
+    resize_transform = transforms.Resize((480, 640))
+    tensor_transforms = transforms.Compose([
+        transforms.ToTensor(),
+        transforms.Normalize([0.485, 0.456, 0.406], [0.229, 0.224, 0.225])
+    ])
     
-image = transforms.Resize((480,640))(Image.fromarray(image))
-tensor = my_transforms(image_aux).unsqueeze(0).to(device)
-
-
-model.to(device)
-with torch.no_grad():
-    outputs = model(tensor)
-
-outputs = torch.argmax(outputs,1)
-
-
-mask = np.array(outputs.cpu())
-mask[mask==0]=255
-mask[mask==1]=150
-mask[mask==2]=76
-mask[mask==3]=25
-mask[mask==4]=0
-
-mask=np.reshape(mask,(480,640))
-return Image.fromarray(mask.astype('uint8'))
-
-
+    # Aplicamos las transformaciones
+    image = resize_transform(Image.fromarray(image))
+    tensor = tensor_transforms(image).unsqueeze(0).to(device)
+
+    # Realizamos la inferencia
+    with torch.no_grad():
+        outputs = model(tensor)
+        outputs = torch.argmax(outputs, 1)
+
+    # Convertimos el tensor de salida a una imagen
+    mask = np.array(outputs.cpu().squeeze(0))
+    mask = np.where(mask == 0, 255, mask)
+    mask = np.where(mask == 1, 150, mask)
+    mask = np.where(mask == 2, 76, mask)
+    mask = np.where(mask == 3, 25, mask)
+    mask = np.where(mask == 4, 0, mask)
     
-# Creamos la interfaz y la lanzamos. 
-gr.Interface(fn=transform_image, inputs=gr.inputs.Image(shape=(640, 480)), outputs=gr.outputs.Image(),examples=['color_154.jpg','color_189.jpg']).launch(share=False)
+    mask = mask.reshape((480, 640))
+    return Image.fromarray(mask.astype('uint8'))
+
+# Creamos la interfaz y la lanzamos.
+gr.Interface(fn=transform_image, inputs=gr.inputs.Image(shape=(640, 480)), outputs=gr.outputs.Image(), examples=['color_154.jpg', 'color_189.jpg']).launch(share=False)