fix

api_service.py

@@ -0,0 +1,49 @@
+import numpy as np
+import cv2
+import onnxruntime as ort
+from fastapi import FastAPI, UploadFile, File
+from fastapi.middleware.cors import CORSMiddleware
+import os
+
+app = FastAPI()
+
+# Add CORS if calling from browser
+app.add_middleware(
+    CORSMiddleware,
+    allow_origins=["*"],
+    allow_methods=["*"],
+    allow_headers=["*"],
+)
+
+# Configuration (same as your Gradio app)
+MODEL_ONNX_PATH = "model.onnx"
+INPUT_SIZE = 640
+CLASS_NAMES = ["class0", "class1"]
+CONF_THRESHOLD = 0.5
+IOU_THRESHOLD = 0.45
+
+# Load ONNX model (from your existing code)
+ort_session = ort.InferenceSession(
+    MODEL_ONNX_PATH,
+    providers=['CUDAExecutionProvider', 'CPUExecutionProvider']
+)
+
+@app.post("/detect")
+async def detect(file: UploadFile = File(...)):
+    # Read image
+    contents = await file.read()
+    nparr = np.frombuffer(contents, np.uint8)
+    image = cv2.imdecode(nparr, cv2.IMREAD_COLOR)
+    
+    # Use your existing predict() function
+    results = predict(image)
+    return results
+
+# Copy your existing predict() and helper functions here
+def predict(image: np.ndarray) -> dict:
+    # ... paste all your processing logic from app.py ...
+    return {"detections": [...]}
+
+if __name__ == "__main__":
+    import uvicorn
+    uvicorn.run(app, host="0.0.0.0", port=8000)

model.onnx

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:39582e24d58748bb0304397f40f5d88bff8f7c7a1584bbec11e9f9490fb12690
+size 37908587

requirements.txt

@@ -0,0 +1,7 @@
+gradio
+numpy
+onnxruntime-gpu
+opencv-python
+Pillow
+torch
+ultralytics