Upload app.py with huggingface_hub

app.py

@@ -2,23 +2,11 @@
 import gradio as gr
 import json
 import traceback
-import os
-
-# Fix PyTorch 2.6 weights_only issue BEFORE importing ultralytics
-import torch
-_orig_load = torch.load
-def _safe_load(*a, **kw):
-    kw.setdefault("weights_only", False)
-    return _orig_load(*a, **kw)
-torch.load = _safe_load
 
 from ultralytics import YOLO
 from huggingface_hub import hf_hub_download
 
-# Download model
 model_path = hf_hub_download(repo_id="foduucom/stockmarket-pattern-detection-yolov8", filename="model.pt")
-print(f"Model path: {model_path}, exists: {os.path.exists(model_path)}, size: {os.path.getsize(model_path)}")
-
 model = YOLO(model_path)
 print(f"Model loaded. Classes: {model.names}")
 
@@ -26,14 +14,9 @@ def detect_patterns(image):
     try:
         if image is None:
             return json.dumps({"patterns": [], "error": "No image"})
-        
-        # Run with verbose to see what happens
-        results = model.predict(source=image, conf=0.20, iou=0.45, imgsz=640, verbose=True)
-        
+        results = model.predict(source=image, conf=0.20, iou=0.45, imgsz=640, verbose=False)
         patterns = []
         for r in results:
-            print(f"Result: boxes={r.boxes.shape if r.boxes is not None else 'None'}, "
-                  f"names={r.names}")
             if r.boxes is None or len(r.boxes) == 0:
                 continue
             for i in range(len(r.boxes)):
@@ -47,14 +30,8 @@ def detect_patterns(image):
                     "confidence": round(conf, 3),
                     "bbox": [round(x, 1) for x in xyxy],
                 })
-        
         patterns.sort(key=lambda p: p["confidence"], reverse=True)
-        return json.dumps({
-            "patterns": patterns,
-            "count": len(patterns),
-            "model_classes": model.names,
-            "image_size": str(image.size) if hasattr(image, 'size') else "unknown",
-        })
+        return json.dumps({"patterns": patterns, "count": len(patterns)})
     except Exception as e:
         return json.dumps({"patterns": [], "error": str(e), "trace": traceback.format_exc()})