Update to API script

app.py

@@ -1,37 +1,32 @@
+import gradio as gr
 from transformers import OwlViTProcessor, OwlViTForObjectDetection
 from PIL import Image
-import torch
-import gradio as gr
 
-# Load model and processor
+# Load the OWL-ViT model and processor
 processor = OwlViTProcessor.from_pretrained("google/owlvit-base-patch32")
 model = OwlViTForObjectDetection.from_pretrained("google/owlvit-base-patch32")
 
+# Define the prediction function
 def predict(image):
-    # Prepare image
-    image = Image.open(image).convert("RGB")
-
-    # Define inputs (zero-shot queries)
-    text_queries = ["A Pokémon", "Pikachu", "Bulbasaur"]
-
-    # Run the model
+    image = Image.open(image)
+    text_queries = ["A photo of a pokemon", "a photo of a human face", "a photo of a couch"]  # Example queries
+    
+    # Prepare inputs for the model
     inputs = processor(text=text_queries, images=image, return_tensors="pt")
-    with torch.no_grad():
-        outputs = model(**inputs)
-
-    # Get predictions
-    target_sizes = torch.tensor([image.size[::-1]])
-    results = processor.post_process_object_detection(outputs, target_sizes=target_sizes, threshold=0.1)
-
-    # Extract boxes
-    boxes = []
-    for score, label, box in zip(results[0]["scores"], results[0]["labels"], results[0]["boxes"]):
-        box = [round(i, 2) for i in box.tolist()]
-        label_text = processor.tokenizer.decode([label])
-        boxes.append({"score": round(score.item(), 3), "label": label_text, "box": box})
+    # Perform inference
+    outputs = model(**inputs)
     
-    return boxes
+    # Format the response (dummy response as example)
+    response = {"message": "Detection successful!"}
+    return response
+
+# Create a Gradio interface and enable the queue (API mode)
+interface = gr.Interface(
+    fn=predict,
+    inputs="image",
+    outputs="json",
+    allow_flagging="never"
+)
 
-# Create Gradio interface
-interface = gr.Interface(fn=predict, inputs="image", outputs="json")
-interface.launch()
+# Launch the interface with API mode enabled
+interface.launch(enable_queue=True)

test.py

@@ -0,0 +1,34 @@
+import requests
+import json
+
+# Replace with your actual Space URL
+API_URL = "https:/shorndrup-owlvit_api.hf.space/predict"
+IMAGE_PATH = r"C:/Users/Administrator/Downloads/pokedex/images/solrock.png"
+
+def call_gradio_api(image_path):
+    # Open the image file in binary mode
+    with open(image_path, "rb") as image_file:
+        files = {"data": (image_path, image_file, "image/png")}
+        response = requests.post(API_URL, files=files)
+
+    # Check for errors
+    if response.status_code != 200:
+        print(f"Error: {response.status_code}")
+        print(response.text)
+        return None
+
+    # Parse the JSON response
+    result = response.json()
+    try:
+        # Extract the predictions from the response
+        predictions = result.get("data", [])[0]
+        if predictions:
+            print("Predictions:", json.dumps(predictions, indent=2))
+        else:
+            print("No predictions found.")
+    except (IndexError, KeyError) as e:
+        print(f"Error parsing response: {e}")
+        print(response.text)
+
+if __name__ == "__main__":
+    call_gradio_api(IMAGE_PATH)