fix(detect): 推理尺寸与训练尺寸不一致时报警

用 640 整图训练的权重跑 160 切片推理，结果看起来完全正常——照常出框、照常有置信度，
但类别系统性地错：实测 mosquito 的 TP 归零，760 张里 599 张被判成 blackdot，
而同一权重整图推理时 mosquito 有 92% 判对。这种错误不会自己暴露。

从 checkpoint 里读训练时的 imgsz，与当前切片尺寸不符就警告。三个调用点都接上：
可视化、批量推送预测、ML backend 加载权重时。visualize 的 summary.json 也补上
tile_size / tile_overlap，免得事后分不清某次产出是怎么跑的。

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>

scripts/visualize_detections.py

@@ -46,7 +46,7 @@ import yaml  # noqa: E402
 
 from datasets.yolo import IMAGE_SUFFIXES  # noqa: E402
 from engine.detection import Box, draw_boxes, match_boxes, parse_label  # noqa: E402
-from engine.tiled_inference import load_image, predict_image  # noqa: E402
+from engine.tiled_inference import check_size_matches, load_image, predict_image  # noqa: E402
 from utils import setup_logging  # noqa: E402
 from utils.device import DeviceNotAvailableError, describe_device, resolve_device  # noqa: E402
 
@@ -180,6 +180,7 @@ def main() -> int:
     from ultralytics import YOLO
 
     model = YOLO(args.weights)
+    check_size_matches(model, args.tile_size, logger)
     device_arg = device.index or 0 if device.type == "cuda" else device.type
 
     outcomes: dict[str, Counter[str]] = {name: Counter() for name in splits}
@@ -246,6 +247,8 @@ def main() -> int:
         "conf": args.conf,
         "iou_threshold": args.iou_threshold,
         "imgsz": args.imgsz,
+        "tile_size": args.tile_size,
+        "tile_overlap": args.tile_overlap,
         "outcomes": {split: dict(counter) for split, counter in outcomes.items()},
         "per_class": {
             split: {