Create app.py

app.py

@@ -0,0 +1,43 @@
+from huggingface_hub import hf_hub_download
+import gradio as gr
+from fastai.vision.all import *
+import torch
+import torchvision.transforms as transforms
+import numpy as np
+from PIL import Image
+
+repo_id = "hafsa101010/Practica3"
+
+model_path = hf_hub_download(repo_id=repo_id, filename="unet.pth")
+
+device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
+model = torch.jit.load(model_path, map_location=device)
+model.eval()
+
+def transform_image(image):
+    image = image.resize((640, 480))  # Asegurar tamaño correcto
+    my_transforms = transforms.Compose([
+        transforms.ToTensor(),
+        transforms.Normalize([0.485, 0.456, 0.406], [0.229, 0.224, 0.225])
+    ])
+    return my_transforms(image).unsqueeze(0).to(device)
+
+def predict(img):
+    tensor = transform_image(img)
+    
+    with torch.no_grad():
+        outputs = model(tensor)
+    
+    outputs = torch.argmax(outputs, 1).cpu().numpy().squeeze()
+
+    mask = np.zeros_like(outputs, dtype=np.uint8)
+    mask[outputs == 1] = 255  # grape
+    mask[outputs == 2] = 150  # leaves
+    mask[outputs == 3] = 76   # pole
+    mask[outputs == 4] = 74   # pole
+    mask[outputs == 5] = 29   # wood
+    mask[outputs == 6] = 25   # wood
+
+    return Image.fromarray(mask)
+
+gr.Interface(fn=predict, inputs=gr.Image(type="pil"), outputs=gr.Image(type="pil"), examples=["color_181.jpg", "color_155.jpg"]).launch()