scorevision: push artifact

miner.py

@@ -1,27 +1,11 @@
-"""miner.py — uploaded to artur7236/turbovision-beverage HF repo.
-
-Fine-tuned YOLO11s on SAM3-labelled beverage data + horizontal-flip TTA at
-1280×1280 + confidence threshold 0.55. Local benchmark mean composite = 0.709
-vs UID 213's 0.667.
-
-Required by the chute template:
-  - class Miner with __init__(self, path_hf_repo: Path)
-  - predict_batch(batch_images, offset, n_keypoints) -> list[TVFrameResult]
-  - BoundingBox + TVFrameResult pydantic models defined in this file
-  - No imports from anywhere except stdlib + site-packages (no helper modules
-    from the HF repo).
-"""
 from pathlib import Path
+import math
 
+import cv2
 import numpy as np
+import onnxruntime as ort
 from numpy import ndarray
 from pydantic import BaseModel
-from ultralytics import YOLO
-
-
-# Class index follows the manifest objects order for Detect-beverage-detect:
-# [cup, bottle, can] → cls_id 0, 1, 2.
-CLASS_NAMES = ["cup", "bottle", "can"]
 
 
 class BoundingBox(BaseModel):
@@ -39,122 +23,478 @@ class TVFrameResult(BaseModel):
     keypoints: list[tuple[int, int]]
 
 
-def _iou(a: BoundingBox, b: BoundingBox) -> float:
-    """Standard IoU between two axis-aligned boxes."""
-    x1 = max(a.x1, b.x1)
-    y1 = max(a.y1, b.y1)
-    x2 = min(a.x2, b.x2)
-    y2 = min(a.y2, b.y2)
-    if x2 <= x1 or y2 <= y1:
-        return 0.0
-    inter = (x2 - x1) * (y2 - y1)
-    area_a = max(0, a.x2 - a.x1) * max(0, a.y2 - a.y1)
-    area_b = max(0, b.x2 - b.x1) * max(0, b.y2 - b.y1)
-    union = area_a + area_b - inter
-    return inter / union if union > 0 else 0.0
-
-
-def _cross_class_nms(boxes: list[BoundingBox], iou_thresh: float = 0.6) -> list[BoundingBox]:
-    """Suppress cross-class overlapping boxes (keep highest-conf when IoU≥thresh).
-
-    Ultralytics' default NMS only dedupes WITHIN a class. The SN44 validator
-    counts cross-class overlapping boxes as false positives (only one class
-    can be right per object), so we need this extra pass. Mirrors UID 213's
-    _cross_class_dedup_op (IoU 0.7); we use 0.6 to be slightly more aggressive.
-    """
-    if len(boxes) <= 1:
-        return boxes
-    sorted_boxes = sorted(boxes, key=lambda b: -b.conf)
-    kept: list[BoundingBox] = []
-    for b in sorted_boxes:
-        suppressed = False
-        for k in kept:
-            if _iou(b, k) >= iou_thresh:
-                suppressed = True
-                break
-        if not suppressed:
-            kept.append(b)
-    return kept
-
-
 class Miner:
-    """Fine-tuned YOLO11s with hflip TTA (Round 5, 607 validator-distribution imgs, 19 MB).
-
-    Tuning rationale (all on SAM3 ground-truth, 361 val images):
-      - imgsz=1280: bumps small-object recall on the typical 1408×768 frames
-        the validator sends (cans, distant cups).
-      - conf=0.55: sweet spot for the validator's pillar weighting
-        (0.6·mAP50 + 0.4·(1−ffpi/10)). Below 0.45 the false-positive pillar
-        crashes; above 0.65 recall drops faster than precision rises.
-      - augment=True: ultralytics enables hflip TTA, our biggest single
-        composite gain in offline sweep.
-      - iou=0.45: standard YOLO NMS IoU; cross-class dedup happens implicitly.
-    """
-
-    IMAGE_SIZE = 1280
-    CONF_THRESH = 0.50  # Round 4: 0.50 narrowly beat 0.45 (0.6982 vs 0.6980)
-    IOU_THRESH = 0.45
-    USE_TTA = True
-    CROSS_CLASS_IOU = 0.6  # Round 6: post-NMS cross-class dedup
-                           # Block 8287800 had 18 boxes including same-pixel cup+bottle pairs.
+    """ONNX Runtime miner. Hard global NMS + sanity filter + dedup + flip TTA, with per-class rescue bonus."""
+
+    class_names = ["cup", "bottle", "can"]
+    input_size = 1280
+    iou_thres = 0.4
+    cross_iou_thresh = 0.7
+    min_side = 8.0
+    min_box_area = 100.0
+    max_aspect_ratio = 10.0
+    max_det = 300
+    _conf_thres_array = np.array([0.6, 0.45, 0.5], dtype=np.float32)
+    _bonus_array = np.array([0.0, 0.0, 0.2], dtype=np.float32)
 
     def __init__(self, path_hf_repo: Path) -> None:
-        weights_path = path_hf_repo / "best.pt"
-        if not weights_path.exists():
-            raise FileNotFoundError(f"missing weights at {weights_path}")
-        self.model = YOLO(str(weights_path))
-        # Touch the model once so cold-start latency hits the warmup, not the
-        # first validator call.
-        dummy = np.zeros((640, 640, 3), dtype=np.uint8)
-        _ = self.model.predict(dummy, imgsz=self.IMAGE_SIZE, conf=self.CONF_THRESH,
-                               iou=self.IOU_THRESH, augment=self.USE_TTA, verbose=False)
-        print(f"✅ YOLO11s loaded from {weights_path}")
+        model_path = path_hf_repo / "weights.onnx"
+        print("ORT version:", ort.__version__)
+
+        try:
+            ort.preload_dlls()
+            print("preload_dlls success")
+        except Exception as e:
+            print(f"preload_dlls failed: {e}")
+
+        print("ORT available providers BEFORE session:", ort.get_available_providers())
+
+        sess_options = ort.SessionOptions()
+        sess_options.graph_optimization_level = ort.GraphOptimizationLevel.ORT_ENABLE_ALL
+
+        try:
+            self.session = ort.InferenceSession(
+                str(model_path),
+                sess_options=sess_options,
+                providers=["CUDAExecutionProvider", "CPUExecutionProvider"],
+            )
+            print("Created ORT session with preferred CUDA provider list")
+        except Exception as e:
+            print(f"CUDA session creation failed, falling back to CPU: {e}")
+            self.session = ort.InferenceSession(
+                str(model_path),
+                sess_options=sess_options,
+                providers=["CPUExecutionProvider"],
+            )
+
+        print("ORT session providers:", self.session.get_providers())
+
+        for inp in self.session.get_inputs():
+            print("INPUT:", inp.name, inp.shape, inp.type)
+        for out in self.session.get_outputs():
+            print("OUTPUT:", out.name, out.shape, out.type)
+
+        self.input_name = self.session.get_inputs()[0].name
+        self.output_names = [output.name for output in self.session.get_outputs()]
+        self.input_shape = self.session.get_inputs()[0].shape
+
+        self.input_height = self._safe_dim(self.input_shape[2], default=self.input_size)
+        self.input_width = self._safe_dim(self.input_shape[3], default=self.input_size)
+
+        print(f"ONNX model loaded from: {model_path}")
+        print(f"ONNX providers: {self.session.get_providers()}")
+        print(f"ONNX input: name={self.input_name}, shape={self.input_shape}")
 
     def __repr__(self) -> str:
-        return (f"YOLO11s_ft(imgsz={self.IMAGE_SIZE}, "
-                f"conf={self.CONF_THRESH}, iou={self.IOU_THRESH}, "
-                f"tta={self.USE_TTA})")
-
-    def predict_batch(
-        self,
-        batch_images: list[ndarray],
-        offset: int,
-        n_keypoints: int,
-    ) -> list[TVFrameResult]:
-        # Run inference. Ultralytics will accept a list[ndarray]; with augment=True
-        # it does the flip+merge internally per image.
-        results = self.model.predict(
-            batch_images,
-            imgsz=self.IMAGE_SIZE,
-            conf=self.CONF_THRESH,
-            iou=self.IOU_THRESH,
-            augment=self.USE_TTA,
-            verbose=False,
+        return (
+            f"ONNXRuntime(session={type(self.session).__name__}, "
+            f"providers={self.session.get_providers()})"
         )
 
-        out: list[TVFrameResult] = []
-        # n_keypoints is irrelevant for detection elements; return zero-padded.
-        kp_zeros = [(0, 0) for _ in range(max(0, int(n_keypoints)))]
-
-        for i, r in enumerate(results):
-            frame_id = offset + i
-            boxes: list[BoundingBox] = []
-            if r.boxes is not None and r.boxes.data is not None:
-                for box in r.boxes.data.cpu().numpy():
-                    x1, y1, x2, y2, conf, cls_id = box.tolist()
-                    cls_id_int = int(cls_id)
-                    if cls_id_int < 0 or cls_id_int >= len(CLASS_NAMES):
-                        continue
-                    # ensure non-degenerate after rounding
-                    xi1, yi1, xi2, yi2 = int(x1), int(y1), int(x2), int(y2)
-                    if xi2 <= xi1 or yi2 <= yi1:
-                        continue
-                    boxes.append(BoundingBox(
-                        x1=xi1, y1=yi1, x2=xi2, y2=yi2,
-                        cls_id=cls_id_int, conf=float(conf),
-                    ))
-            # Cross-class dedup (Ultralytics NMS is per-class only)
-            boxes = _cross_class_nms(boxes, iou_thresh=self.CROSS_CLASS_IOU)
-            out.append(TVFrameResult(frame_id=frame_id, boxes=boxes, keypoints=kp_zeros))
+    @staticmethod
+    def _safe_dim(value, default: int) -> int:
+        return value if isinstance(value, int) and value > 0 else default
+
+    def _letterbox(self, image: ndarray, new_shape: tuple[int, int],
+                   color=(114, 114, 114)
+                   ) -> tuple[ndarray, float, tuple[float, float]]:
+        h, w = image.shape[:2]
+        new_w, new_h = new_shape
+        ratio = min(new_w / w, new_h / h)
+        resized_w = int(round(w * ratio))
+        resized_h = int(round(h * ratio))
+        if (resized_w, resized_h) != (w, h):
+            interp = cv2.INTER_CUBIC if ratio > 1.0 else cv2.INTER_LINEAR
+            image = cv2.resize(image, (resized_w, resized_h), interpolation=interp)
+        dw = (new_w - resized_w) / 2.0
+        dh = (new_h - resized_h) / 2.0
+        left = int(round(dw - 0.1))
+        right = int(round(dw + 0.1))
+        top = int(round(dh - 0.1))
+        bottom = int(round(dh + 0.1))
+        padded = cv2.copyMakeBorder(image, top, bottom, left, right,
+                                    borderType=cv2.BORDER_CONSTANT, value=color)
+        return padded, ratio, (dw, dh)
+
+    def _preprocess(self, image: ndarray
+                    ) -> tuple[np.ndarray, float, tuple[float, float],
+                               tuple[int, int]]:
+        orig_h, orig_w = image.shape[:2]
+        img, ratio, pad = self._letterbox(image, (self.input_width, self.input_height))
+        img = cv2.cvtColor(img, cv2.COLOR_BGR2RGB)
+        img = img.astype(np.float32) / 255.0
+        img = np.transpose(img, (2, 0, 1))[None, ...]
+        img = np.ascontiguousarray(img, dtype=np.float32)
+        return img, ratio, pad, (orig_w, orig_h)
+
+    @staticmethod
+    def _clip_boxes(boxes: np.ndarray, image_size: tuple[int, int]) -> np.ndarray:
+        w, h = image_size
+        boxes[:, 0] = np.clip(boxes[:, 0], 0, w - 1)
+        boxes[:, 1] = np.clip(boxes[:, 1], 0, h - 1)
+        boxes[:, 2] = np.clip(boxes[:, 2], 0, w - 1)
+        boxes[:, 3] = np.clip(boxes[:, 3], 0, h - 1)
+        return boxes
+
+    @staticmethod
+    def _xywh_to_xyxy(boxes: np.ndarray) -> np.ndarray:
+        out = np.empty_like(boxes)
+        out[:, 0] = boxes[:, 0] - boxes[:, 2] / 2.0
+        out[:, 1] = boxes[:, 1] - boxes[:, 3] / 2.0
+        out[:, 2] = boxes[:, 0] + boxes[:, 2] / 2.0
+        out[:, 3] = boxes[:, 1] + boxes[:, 3] / 2.0
+        return out
+
+    @staticmethod
+    def _hard_nms(boxes: np.ndarray, scores: np.ndarray,
+                  iou_thresh: float) -> np.ndarray:
+        n = len(boxes)
+        if n == 0:
+            return np.array([], dtype=np.intp)
+        order = np.argsort(-scores)
+        keep: list[int] = []
+        while len(order) > 0:
+            i = int(order[0])
+            keep.append(i)
+            if len(order) == 1:
+                break
+            rest = order[1:]
+            xx1 = np.maximum(boxes[i, 0], boxes[rest, 0])
+            yy1 = np.maximum(boxes[i, 1], boxes[rest, 1])
+            xx2 = np.minimum(boxes[i, 2], boxes[rest, 2])
+            yy2 = np.minimum(boxes[i, 3], boxes[rest, 3])
+            inter = np.maximum(0.0, xx2 - xx1) * np.maximum(0.0, yy2 - yy1)
+            a_i = (max(0.0, boxes[i, 2] - boxes[i, 0]) *
+                   max(0.0, boxes[i, 3] - boxes[i, 1]))
+            a_r = (np.maximum(0.0, boxes[rest, 2] - boxes[rest, 0]) *
+                   np.maximum(0.0, boxes[rest, 3] - boxes[rest, 1]))
+            iou = inter / (a_i + a_r - inter + 1e-7)
+            order = rest[iou <= iou_thresh]
+        return np.array(keep, dtype=np.intp)
+
+    def _per_class_hard_nms(self, boxes: np.ndarray, scores: np.ndarray,
+                            cls_ids: np.ndarray, iou_thresh: float
+                            ) -> np.ndarray:
+        if len(boxes) == 0:
+            return np.array([], dtype=np.intp)
+        all_keep: list[int] = []
+        for c in np.unique(cls_ids):
+            mask = cls_ids == c
+            indices = np.where(mask)[0]
+            keep = self._hard_nms(boxes[mask], scores[mask], iou_thresh)
+            all_keep.extend(indices[keep].tolist())
+        all_keep.sort()
+        return np.array(all_keep, dtype=np.intp)
+
+    def _cross_class_dedup_op(self, boxes: np.ndarray, scores: np.ndarray,
+                              cls_ids: np.ndarray, iou_thresh: float
+                              ) -> tuple[np.ndarray, np.ndarray, np.ndarray]:
+        n = len(boxes)
+        if n <= 1:
+            return boxes, scores, cls_ids
+        boxes = np.asarray(boxes, dtype=np.float32)
+        scores = np.asarray(scores, dtype=np.float32)
+        cls_ids = np.asarray(cls_ids, dtype=np.int32)
+        areas = (np.maximum(0.0, boxes[:, 2] - boxes[:, 0]) *
+                 np.maximum(0.0, boxes[:, 3] - boxes[:, 1]))
+        margins = scores - self._conf_thres_array[cls_ids]
+        order = np.lexsort((-areas, -margins))
+        suppressed = np.zeros(n, dtype=bool)
+        keep: list[int] = []
+        for i in order:
+            if suppressed[i]:
+                continue
+            keep.append(int(i))
+            bi = boxes[i]
+            xx1 = np.maximum(bi[0], boxes[:, 0])
+            yy1 = np.maximum(bi[1], boxes[:, 1])
+            xx2 = np.minimum(bi[2], boxes[:, 2])
+            yy2 = np.minimum(bi[3], boxes[:, 3])
+            inter = np.maximum(0.0, xx2 - xx1) * np.maximum(0.0, yy2 - yy1)
+            a_i = max(1e-7, float((bi[2] - bi[0]) * (bi[3] - bi[1])))
+            iou = inter / (a_i + areas - inter + 1e-7)
+            dup = iou > iou_thresh
+            dup[i] = False
+            suppressed |= dup
+        keep_idx = np.array(keep, dtype=np.intp)
+        return boxes[keep_idx], scores[keep_idx], cls_ids[keep_idx]
+
+    def _filter_sane_boxes(self, boxes: np.ndarray, scores: np.ndarray,
+                           cls_ids: np.ndarray, orig_size: tuple[int, int]
+                           ) -> tuple[np.ndarray, np.ndarray, np.ndarray]:
+        if len(boxes) == 0:
+            return boxes, scores, cls_ids
+        orig_w, orig_h = orig_size
+        image_area = float(orig_w * orig_h)
+        bw = np.maximum(0.0, boxes[:, 2] - boxes[:, 0])
+        bh = np.maximum(0.0, boxes[:, 3] - boxes[:, 1])
+        area = bw * bh
+        ar = np.where(
+            (bw > 0) & (bh > 0),
+            np.maximum(bw / np.maximum(bh, 1e-6), bh / np.maximum(bw, 1e-6)),
+            np.inf,
+        )
+        keep = (
+            (bw >= self.min_side) & (bh >= self.min_side) &
+            (area >= self.min_box_area) &
+            (area <= 0.95 * image_area) &
+            (ar <= self.max_aspect_ratio)
+        )
+        return boxes[keep], scores[keep], cls_ids[keep]
 
+    def _max_score_per_cluster(self, post_boxes: np.ndarray,
+                               post_cls: np.ndarray,
+                               full_boxes: np.ndarray,
+                               full_scores: np.ndarray,
+                               full_cls: np.ndarray,
+                               iou_thresh: float) -> np.ndarray:
+        n = len(post_boxes)
+        if n == 0:
+            return np.empty(0, dtype=np.float32)
+        full_areas = (np.maximum(0.0, full_boxes[:, 2] - full_boxes[:, 0]) *
+                      np.maximum(0.0, full_boxes[:, 3] - full_boxes[:, 1]))
+        out = np.empty(n, dtype=np.float32)
+        for i in range(n):
+            bi = post_boxes[i]
+            xx1 = np.maximum(bi[0], full_boxes[:, 0])
+            yy1 = np.maximum(bi[1], full_boxes[:, 1])
+            xx2 = np.minimum(bi[2], full_boxes[:, 2])
+            yy2 = np.minimum(bi[3], full_boxes[:, 3])
+            inter = np.maximum(0.0, xx2 - xx1) * np.maximum(0.0, yy2 - yy1)
+            a_i = max(0.0, float((bi[2] - bi[0]) * (bi[3] - bi[1])))
+            iou = inter / (a_i + full_areas - inter + 1e-7)
+            cluster = (iou >= iou_thresh) & (full_cls == post_cls[i])
+            out[i] = float(np.max(full_scores[cluster])) if np.any(cluster) else 0.0
         return out
+
+    def _conf_filter_mask(self, scores: np.ndarray,
+                          cls_ids: np.ndarray) -> np.ndarray:
+        """Boolean keep-mask: score >= per-class threshold, with a per-class
+        rescue — if a class has zero boxes passing, admit its top-1 candidate
+        when its score >= (per-class threshold - per-class bonus)."""
+        if len(scores) == 0:
+            return np.zeros(0, dtype=bool)
+        thr = self._conf_thres_array[cls_ids]
+        keep = scores >= thr
+        for c in np.unique(cls_ids):
+            b = float(self._bonus_array[c])
+            if b <= 0.0:
+                continue
+            cm = cls_ids == c
+            if keep[cm].any():
+                continue
+            idx = np.where(cm)[0]
+            top = int(idx[int(np.argmax(scores[idx]))])
+            if scores[top] >= self._conf_thres_array[c] - b:
+                keep[top] = True
+        return keep
+
+    def _per_view_pipeline(self, boxes: np.ndarray, scores: np.ndarray,
+                           cls_ids: np.ndarray, orig_size: tuple[int, int]
+                           ) -> tuple[np.ndarray, np.ndarray, np.ndarray]:
+        boxes, scores, cls_ids = self._filter_sane_boxes(
+            boxes, scores, cls_ids, orig_size
+        )
+        if len(boxes) == 0:
+            return boxes, scores, cls_ids
+        if len(boxes) > 1:
+            keep = self._hard_nms(boxes, scores, self.iou_thres)
+            boxes, scores, cls_ids = boxes[keep], scores[keep], cls_ids[keep]
+        if len(scores) > self.max_det:
+            top = np.argsort(-scores)[: self.max_det]
+            boxes, scores, cls_ids = boxes[top], scores[top], cls_ids[top]
+        if len(boxes) > 1:
+            boxes, scores, cls_ids = self._cross_class_dedup_op(
+                boxes, scores, cls_ids, self.cross_iou_thresh
+            )
+        return boxes, scores, cls_ids
+
+    def _decode_final_dets(self, preds: np.ndarray, ratio: float,
+                           pad: tuple[float, float],
+                           orig_size: tuple[int, int]) -> list[BoundingBox]:
+        if preds.ndim == 3 and preds.shape[0] == 1:
+            preds = preds[0]
+        if preds.ndim != 2 or preds.shape[1] < 6:
+            raise ValueError(f"Unexpected ONNX final-det output shape: {preds.shape}")
+
+        boxes = preds[:, :4].astype(np.float32)
+        scores = preds[:, 4].astype(np.float32)
+        cls_ids = preds[:, 5].astype(np.int32)
+
+        keep = self._conf_filter_mask(scores, cls_ids)
+        boxes = boxes[keep]
+        scores = scores[keep]
+        cls_ids = cls_ids[keep]
+        if len(boxes) == 0:
+            return []
+
+        pad_w, pad_h = pad
+        boxes[:, [0, 2]] -= pad_w
+        boxes[:, [1, 3]] -= pad_h
+        boxes /= ratio
+        boxes = self._clip_boxes(boxes, orig_size)
+
+        boxes, scores, cls_ids = self._per_view_pipeline(
+            boxes, scores, cls_ids, orig_size
+        )
+        return self._build_results(boxes, scores, cls_ids)
+
+    def _decode_raw_yolo(self, preds: np.ndarray, ratio: float,
+                         pad: tuple[float, float],
+                         orig_size: tuple[int, int]) -> list[BoundingBox]:
+        if preds.ndim != 3 or preds.shape[0] != 1:
+            raise ValueError(f"Unexpected raw ONNX output shape: {preds.shape}")
+        preds = preds[0]
+        if preds.shape[0] <= 16 and preds.shape[1] > preds.shape[0]:
+            preds = preds.T
+        if preds.ndim != 2 or preds.shape[1] < 5:
+            raise ValueError(f"Unexpected raw output shape: {preds.shape}")
+
+        boxes_xywh = preds[:, :4].astype(np.float32)
+        cls_part = preds[:, 4:].astype(np.float32)
+        if cls_part.shape[1] == 1:
+            scores = cls_part[:, 0]
+            cls_ids = np.zeros(len(scores), dtype=np.int32)
+        else:
+            cls_ids = np.argmax(cls_part, axis=1).astype(np.int32)
+            scores = cls_part[np.arange(len(cls_part)), cls_ids]
+
+        keep = self._conf_filter_mask(scores, cls_ids)
+        boxes_xywh = boxes_xywh[keep]
+        scores = scores[keep]
+        cls_ids = cls_ids[keep]
+        if len(boxes_xywh) == 0:
+            return []
+        boxes = self._xywh_to_xyxy(boxes_xywh)
+
+        pad_w, pad_h = pad
+        boxes[:, [0, 2]] -= pad_w
+        boxes[:, [1, 3]] -= pad_h
+        boxes /= ratio
+        boxes = self._clip_boxes(boxes, orig_size)
+
+        boxes, scores, cls_ids = self._per_view_pipeline(
+            boxes, scores, cls_ids, orig_size
+        )
+        return self._build_results(boxes, scores, cls_ids)
+
+    @staticmethod
+    def _build_results(boxes: np.ndarray, scores: np.ndarray,
+                       cls_ids: np.ndarray) -> list[BoundingBox]:
+        results: list[BoundingBox] = []
+        for box, conf, cls_id in zip(boxes, scores, cls_ids):
+            x1, y1, x2, y2 = box.tolist()
+            if x2 <= x1 or y2 <= y1:
+                continue
+            results.append(
+                BoundingBox(
+                    x1=int(math.floor(x1)),
+                    y1=int(math.floor(y1)),
+                    x2=int(math.ceil(x2)),
+                    y2=int(math.ceil(y2)),
+                    cls_id=int(cls_id),
+                    conf=float(conf),
+                )
+            )
+        return results
+
+    def _postprocess(self, output: np.ndarray, ratio: float,
+                     pad: tuple[float, float],
+                     orig_size: tuple[int, int]) -> list[BoundingBox]:
+        if output.ndim == 2 and output.shape[1] >= 6:
+            return self._decode_final_dets(output, ratio, pad, orig_size)
+        if output.ndim == 3 and output.shape[0] == 1 and output.shape[2] == 6:
+            return self._decode_final_dets(output, ratio, pad, orig_size)
+        return self._decode_raw_yolo(output, ratio, pad, orig_size)
+
+    def _predict_single(self, image: np.ndarray) -> list[BoundingBox]:
+        if image is None:
+            raise ValueError("Input image is None")
+        if not isinstance(image, np.ndarray):
+            raise TypeError(f"Input is not numpy array: {type(image)}")
+        if image.ndim != 3:
+            raise ValueError(f"Expected HWC image, got shape={image.shape}")
+        if image.shape[2] != 3:
+            raise ValueError(f"Expected 3 channels, got shape={image.shape}")
+        if image.dtype != np.uint8:
+            image = image.astype(np.uint8)
+
+        input_tensor, ratio, pad, orig_size = self._preprocess(image)
+        expected = (1, 3, self.input_height, self.input_width)
+        if input_tensor.shape != expected:
+            raise ValueError(
+                f"Bad input tensor shape={input_tensor.shape}, expected={expected}"
+            )
+
+        outputs = self.session.run(self.output_names, {self.input_name: input_tensor})
+        return self._postprocess(outputs[0], ratio, pad, orig_size)
+
+    def _predict_tta(self, image: np.ndarray) -> list[BoundingBox]:
+        boxes_orig = self._predict_single(image)
+        flipped = cv2.flip(image, 1)
+        boxes_flip = self._predict_single(flipped)
+        w = image.shape[1]
+        boxes_flip = [
+            BoundingBox(
+                x1=w - b.x2, y1=b.y1, x2=w - b.x1, y2=b.y2,
+                cls_id=b.cls_id, conf=b.conf,
+            )
+            for b in boxes_flip
+        ]
+        all_boxes = boxes_orig + boxes_flip
+        if not all_boxes:
+            return []
+
+        coords = np.array(
+            [[b.x1, b.y1, b.x2, b.y2] for b in all_boxes], dtype=np.float32
+        )
+        scores = np.array([b.conf for b in all_boxes], dtype=np.float32)
+        cls_ids = np.array([b.cls_id for b in all_boxes], dtype=np.int32)
+
+        hard_keep = self._per_class_hard_nms(coords, scores, cls_ids, self.iou_thres)
+        if len(hard_keep) == 0:
+            return []
+        if len(hard_keep) > self.max_det:
+            top = np.argsort(-scores[hard_keep])[: self.max_det]
+            hard_keep = hard_keep[top]
+        boosted = self._max_score_per_cluster(
+            coords[hard_keep], cls_ids[hard_keep],
+            coords, scores, cls_ids, self.iou_thres,
+        )
+
+        kept_coords = coords[hard_keep]
+        kept_cls = cls_ids[hard_keep]
+        if len(kept_coords) > 1:
+            kept_coords, boosted, kept_cls = self._cross_class_dedup_op(
+                kept_coords, boosted, kept_cls, self.cross_iou_thresh
+            )
+
+        return [
+            BoundingBox(
+                x1=int(math.floor(kept_coords[j, 0])),
+                y1=int(math.floor(kept_coords[j, 1])),
+                x2=int(math.ceil(kept_coords[j, 2])),
+                y2=int(math.ceil(kept_coords[j, 3])),
+                cls_id=int(kept_cls[j]),
+                conf=float(boosted[j]),
+            )
+            for j in range(len(kept_coords))
+        ]
+
+    def predict_batch(self, batch_images: list[ndarray], offset: int,
+                      n_keypoints: int) -> list[TVFrameResult]:
+        results: list[TVFrameResult] = []
+        for frame_number_in_batch, image in enumerate(batch_images):
+            try:
+                boxes = self._predict_tta(image)
+            except Exception as e:
+                print(f"Inference failed for frame {offset + frame_number_in_batch}: {e}")
+                boxes = []
+            results.append(
+                TVFrameResult(
+                    frame_id=offset + frame_number_in_batch,
+                    boxes=boxes,
+                    keypoints=[(0, 0) for _ in range(max(0, int(n_keypoints)))],
+                )
+            )
+        return results