Update app.py

app.py

@@ -1,17 +1,17 @@
 import gradio as gr
 from transformers import pipeline
 
-# ViT-Modell (dein trainiertes Modell von Hugging Face)
+# ViT-Modell (dein trainiertes Modell)
 vit_classifier = pipeline(
     "image-classification",
     model="dewiri/vit-base-oxford-iiit-pets",
     top_k=3,
-    device=0  # GPU, falls verfügbar
+    device=0
 )
 
-# SIGLIP für Zero-Shot Klassifikation
-siglip_detector = pipeline(
-    model="google/siglip-so400m-patch14-384",
+# CLIP-Modell für Zero-Shot Klassifikation
+clip_detector = pipeline(
+    model="openai/clip-vit-large-patch14",
     task="zero-shot-image-classification",
     top_k=3,
     device=0
@@ -29,12 +29,12 @@ def classify_pet(image):
     vit_results = vit_classifier(image)
     vit_output = {res['label']: round(res['score'], 3) for res in vit_results}
     
-    siglip_results = siglip_detector(image, candidate_labels=labels_oxford_pets)
-    siglip_output = {res['label']: round(res['score'], 3) for res in siglip_results}
+    clip_results = clip_detector(image, candidate_labels=labels_oxford_pets)
+    clip_output = {res['label']: round(res['score'], 3) for res in clip_results}
     
     return {
         "ViT Classification (dewiri)": vit_output,
-        "SIGLIP Zero-Shot Classification": siglip_output
+        "CLIP Zero-Shot Classification": clip_output
     }
 
 example_images = [
@@ -50,7 +50,7 @@ iface = gr.Interface(
     inputs=gr.Image(type="filepath"),
     outputs=gr.JSON(),
     title="Pet Classification Comparison",
-    description="Compare a fine-tuned ViT model (dewiri/vit-base-oxford-iiit-pets) with a SIGLIP zero-shot classifier.",
+    description="Compare a fine-tuned ViT model (dewiri/vit-base-oxford-iiit-pets) with OpenAI's CLIP zero-shot classifier.",
     examples=example_images
 )