fix3: torch.load

app.py

@@ -48,6 +48,18 @@ def _patch_gradio_client():
         pass
 _patch_gradio_client()
 
+# ── PyTorch 2.7 global weights_only patch ────────────────────────────────────
+# YOLO and other libraries call torch.load internally without weights_only=False.
+# Monkey-patch torch.load to default to weights_only=False for all calls.
+import torch as _torch
+_orig_torch_load = _torch.load
+def _patched_torch_load(f, map_location=None, pickle_module=None,
+                         weights_only=False, mmap=None, **kwargs):
+    return _orig_torch_load(f, map_location=map_location,
+                             pickle_module=pickle_module,
+                             weights_only=False, **kwargs)
+_torch.load = _patched_torch_load
+
 import os
 import sys
 import json
@@ -799,9 +811,9 @@ def run_yolo_seg(frame_rgb: np.ndarray, seg_conf: float) -> np.ndarray:
         n_written += int(binary.sum())
 
     detected_cls = [_YOLO_CLASS_NAMES[min(int(c) + 1, 25)] for c in cls_ids]
-    print(f"[YOLO] Detections: {len(cls_ids)}  |  classes: {detected_cls}  |  "
+    print(f"Segmentation Detections: {len(cls_ids)}  |  classes: {detected_cls}  |  "
           f"vessel px: {n_written} ({100 * n_written / (h * w):.1f}%)")
-    print(f"[YOLO] seg_map unique labels: {np.unique(seg_map)}")
+    print(f"seg_map unique labels: {np.unique(seg_map)}")
     return seg_map
 
 # ─────────────────────────────────────────────────────────────────────────────
@@ -1338,7 +1350,7 @@ def analyse(video_path, sten_conf, seg_conf, px_per_mm_override, progress=gr.Pro
             det["overlap"]=float(binary_mask[y1:y2,x1:x2].sum())/max(area,1)
 
         # Step 3b — YOLOv8m-seg 26-class segmentation
-        progress(0.65, desc="Running YOLOv8m-seg…")
+        progress(0.65, desc="Running seg model")
         seg_map = run_yolo_seg(frame_rgb, seg_conf)
         seg_overlay = render_nnunet_overlay(frame_rgb, seg_map)
 
@@ -1498,7 +1510,7 @@ with gr.Blocks(css=CSS, title="Angio AI") as demo:
                     )
                     seg_conf = gr.Slider(
                         minimum=0.05, maximum=0.95, value=0.25, step=0.05,
-                        label="Segmentation confidence threshold ",
+                        label="Segmentation confidence threshold",
                         info="Detections below this confidence are discarded by NMS",
                     )
                     px_per_mm = gr.Slider(