Update app.py

app.py

@@ -4,6 +4,7 @@ import torch
 from torchvision import transforms
 from PIL import Image
 import io
+from huggingface_hub import hf_hub_download
 
 # FastAPI uygulamasını başlat
 app = FastAPI()
@@ -11,8 +12,13 @@ app = FastAPI()
 # Cihaz ayarı
 device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
 
-# Eğitilmiş modeli yükleme
-model = torch.load("best_model.pth", map_location=device)
+# Hugging Face modelini yükleme
+REPO_ID = "kedimestan/retinoblastomaDetectionVGG19"
+MODEL_FILE = "best_model.bin"
+
+# Model dosyasını indirin
+model_path = hf_hub_download(repo_id=REPO_ID, filename=MODEL_FILE)
+model = torch.load(model_path, map_location=device)
 model.eval()
 
 # Görüntü dönüşüm pipeline'ı
@@ -26,9 +32,9 @@ transform = transforms.Compose([
 def predict(image: Image.Image):
     input_tensor = transform(image).unsqueeze(0).to(device)
     with torch.no_grad():
-        output = model(input_tensor).item()  # Tahmini al
-        prediction = "Positive" if output > 0.5 else "Negative"
-        return {"Prediction": prediction, "Probability": round(output, 2)}
+        output = model(input_tensor).squeeze(0).cpu().numpy()
+        prediction = "Positive" if output[0] > 0.5 else "Negative"
+        return {"Prediction": prediction, "Probability": round(float(output[0]), 2)}
 
 # Ana API rotası
 @app.post("/predict")
@@ -47,4 +53,4 @@ async def predict_image(file: UploadFile = File(...)):
 # Ana sayfa
 @app.get("/")
 def home():
-    return {"message": "Upload an image to /predict for classification."}
+    return {"message": "Upload an image to /predict for classification."}