@camstack/addon-pipeline 1.2.294 → 1.2.295

package/python/postprocessors/test_ssd.py

@@ -0,0 +1,205 @@
+"""Pins the SSD decoder on REAL SSDLite MobileDet output tensors.
+
+`ssdlite-mobiledet-coco-edgetpu` joins the two Coral detectors that already use
+`postprocessor: 'ssd'`. Its graph ends in the same `TFLite_Detection_PostProcess`
+op, but nothing had checked that the decoder reads MobileDet's outputs the way
+the model writes them. Each class below pins one thing that could silently turn
+a working model into wrong labels or empty frames:
+
+- the output ORDER (boxes, classes, scores, count on positional keys "0".."3");
+- the class-index OFFSET (91 score columns, background dropped by the op, so
+  the emitted index is 0-based and 0 is person);
+- the COUNT clamp (MobileDet emits N=100 rows, not 20 like SSD MobileNet V2);
+- the letterbox inversion, against boxes an unrelated detector (DETR) drew on
+  the same COCO frame;
+- the score CEILING of this quantised graph (about 0.771), which decides which
+  confidence thresholds can ever pass.
+
+The fixture (`testdata/ssdlite_mobiledet_outputs.json`) holds the model's raw
+outputs on two COCO val2017 frames; its `_provenance` says how they were made.
+
+`unittest.TestCase` with no numpy: `scripts/test-python-pool.sh` runs these on
+CI's bare interpreter, where numpy is absent, and `ssd.py` decodes plain lists
+for exactly that reason. The one case that needs numpy (ndarray inputs decode
+the same as lists) is skipped, by name, without it.
+"""
+import json
+import os
+import unittest
+
+try:  # real numpy when the sandbox has it
+    import numpy as _numpy_probe
+except ImportError:
+    _numpy_probe = None
+
+# `ndarray` is real-numpy-only: a stub another test module installed would import.
+_HAS_REAL_NUMPY = _numpy_probe is not None and hasattr(_numpy_probe, "ndarray")
+
+from ssd import COCO_90, label_for_class, postprocess_ssd  # noqa: E402
+
+_FIXTURE = os.path.join(os.path.dirname(os.path.abspath(__file__)), "testdata",
+                        "ssdlite_mobiledet_outputs.json")
+with open(_FIXTURE, encoding="utf-8") as _fh:
+    _FRAMES = json.load(_fh)["frames"]
+
+CATS = _FRAMES["coco_val2017_000000039769_two_cats"]
+STREET = _FRAMES["coco_val2017_000000252219_street"]
+
+# The catalog entry's input side; the pool sets `inputSize` from the graph.
+INPUT_SIZE = 320
+# The score column of this graph is LOGISTIC over a uint8 logit quantised at
+# scale 0.046695, zero point 229, so the largest logit it can carry is
+# (255 - 229) * 0.046695 = 1.2141 and sigmoid(1.2141) = 0.77102.
+SCORE_CEILING = 0.77102
+
+
+def _decode(frame: dict, confidence: float = 0.5, outputs: "dict | None" = None) -> list:
+    width, height = frame["orig"]
+    result = postprocess_ssd(
+        outputs if outputs is not None else frame["outputs"],
+        {"confidence": confidence, "inputSize": INPUT_SIZE},
+        width, height, frame["scale"], tuple(frame["pad"]),
+    )
+    return result["detections"]
+
+
+def _iou(a: list, b: list) -> float:
+    ix = max(0.0, min(a[2], b[2]) - max(a[0], b[0]))
+    iy = max(0.0, min(a[3], b[3]) - max(a[1], b[1]))
+    inter = ix * iy
+    union = (a[2] - a[0]) * (a[3] - a[1]) + (b[2] - b[0]) * (b[3] - b[1]) - inter
+    return inter / union if union > 0 else 0.0
+
+
+class FixtureShapeTest(unittest.TestCase):
+    """The fixture is what the MobileDet graph emits: [1,100,4] [1,100] [1,100] [1]."""
+
+    def test_output_shapes(self) -> None:
+        for frame in (CATS, STREET):
+            out = frame["outputs"]
+            self.assertEqual(len(out["0"]), 1)
+            self.assertEqual(len(out["0"][0]), 100)
+            self.assertTrue(all(len(row) == 4 for row in out["0"][0]))
+            self.assertEqual(len(out["1"][0]), 100)
+            self.assertEqual(len(out["2"][0]), 100)
+            self.assertEqual(out["3"], [100.0])
+
+    def test_scores_are_sorted_descending(self) -> None:
+        # TFLite_Detection_PostProcess ranks by score; a reordered export would not.
+        scores = STREET["outputs"]["2"][0]
+        self.assertEqual(scores, sorted(scores, reverse=True))
+
+
+class OutputOrderTest(unittest.TestCase):
+    """Positional "1" is CLASSES and "2" is SCORES. Swapped, the labels are nonsense."""
+
+    def test_cats_decode_as_cats(self) -> None:
+        labels = [d["class"] for d in _decode(CATS)]
+        self.assertEqual(labels.count("cat"), 2, labels)
+
+    def test_street_decodes_three_people(self) -> None:
+        people = [d for d in _decode(STREET) if d["class"] == "person"]
+        self.assertEqual(len(people), 3)
+        # Left, middle and right of the frame.
+        centres = sorted((d["bbox"][0] + d["bbox"][2]) / 2 for d in people)
+        self.assertLess(centres[0], 200)
+        self.assertTrue(250 < centres[1] < 450, centres)
+        self.assertGreater(centres[2], 500)
+
+    def test_every_score_is_a_probability(self) -> None:
+        for frame in (CATS, STREET):
+            self.assertTrue(all(0.0 <= s <= 1.0 for s in frame["outputs"]["2"][0]))
+
+
+class ClassOffsetTest(unittest.TestCase):
+    """Emitted class index = COCO-90 line number of `coco_labels.txt`, 0 = person."""
+
+    def test_raw_indices_behind_the_labels(self) -> None:
+        classes = STREET["outputs"]["1"][0]
+        scores = STREET["outputs"]["2"][0]
+        top = [int(c) for c, s in zip(classes, scores) if s >= 0.7]
+        self.assertEqual(set(top), {0})  # the three people
+        cat_classes = {int(c) for c, s in zip(CATS["outputs"]["1"][0], CATS["outputs"]["2"][0]) if s >= 0.7}
+        self.assertEqual(cat_classes, {16})
+
+    def test_label_map(self) -> None:
+        self.assertEqual(len(COCO_90), 90)
+        self.assertEqual(label_for_class(0), "person")
+        self.assertEqual(label_for_class(2), "car")
+        self.assertEqual(label_for_class(16), "cat")
+        self.assertEqual(label_for_class(17), "dog")
+        self.assertEqual(label_for_class(89), "toothbrush")
+        self.assertEqual(label_for_class(11), "n/a")
+        self.assertEqual(label_for_class(90), "90")
+
+    def test_gap_classes_are_dropped(self) -> None:
+        out = {k: v for k, v in STREET["outputs"].items()}
+        out["1"] = [[11.0] + list(STREET["outputs"]["1"][0][1:])]  # "n/a" on the top row
+        labels = [d["class"] for d in _decode(STREET, outputs=out)]
+        self.assertNotIn("n/a", labels)
+        self.assertEqual(labels.count("person"), 2)
+
+
+class CountClampTest(unittest.TestCase):
+    """All 100 rows are read when count says 100, and only `count` rows otherwise."""
+
+    def test_low_threshold_reads_past_row_20(self) -> None:
+        # SSD MobileNet V2 stops at 20 rows; MobileDet's detections go on to row 100.
+        dets = _decode(STREET, confidence=0.01)
+        self.assertGreater(len(dets), 20)
+
+    def test_count_limits_the_rows(self) -> None:
+        out = dict(STREET["outputs"])
+        out["3"] = [2.0]
+        self.assertEqual(len(_decode(STREET, confidence=0.01, outputs=out)), 2)
+
+
+class LetterboxGeometryTest(unittest.TestCase):
+    """Boxes land in ORIGINAL-frame pixels, where another detector puts the same cats."""
+
+    # DETR ResNet-50 on this exact COCO frame (the facebook/detr-resnet-50 model
+    # card example): an independent reference for where the two cats are.
+    DETR_CATS = ([13.24, 52.05, 314.02, 470.93], [345.4, 23.85, 640.37, 368.72])
+
+    def test_cats_match_the_reference_boxes(self) -> None:
+        cats = [d["bbox"] for d in _decode(CATS) if d["class"] == "cat"]
+        for ref in self.DETR_CATS:
+            best = max(_iou(ref, box) for box in cats)
+            self.assertGreater(best, 0.85, (ref, cats))
+
+    def test_boxes_stay_inside_the_frame(self) -> None:
+        for frame in (CATS, STREET):
+            width, height = frame["orig"]
+            for d in _decode(frame, confidence=0.01):
+                x1, y1, x2, y2 = d["bbox"]
+                self.assertTrue(0 <= x1 < x2 <= width and 0 <= y1 < y2 <= height, d)
+
+
+class ScoreCeilingTest(unittest.TestCase):
+    """This graph cannot score above ~0.771, so a stricter threshold returns nothing."""
+
+    def test_no_score_exceeds_the_ceiling(self) -> None:
+        for frame in (CATS, STREET):
+            self.assertLessEqual(max(frame["outputs"]["2"][0]), SCORE_CEILING)
+
+    def test_the_ceiling_is_reached(self) -> None:
+        self.assertAlmostEqual(max(CATS["outputs"]["2"][0]), 0.76953125)
+
+    def test_threshold_above_the_ceiling_finds_nothing(self) -> None:
+        self.assertEqual(_decode(STREET, confidence=0.78), [])
+        self.assertEqual(len([d for d in _decode(STREET, confidence=0.65) if d["class"] == "person"]), 3)
+
+
+class NumpyInputTest(unittest.TestCase):
+    """The pool hands ndarrays; they must decode exactly like the lists above."""
+
+    @unittest.skipUnless(_HAS_REAL_NUMPY, "needs real numpy (not on CI's bare interpreter)")
+    def test_ndarrays_decode_like_lists(self) -> None:
+        np = _numpy_probe
+        for frame in (CATS, STREET):
+            arrays = {k: np.asarray(v, dtype=np.float32) for k, v in frame["outputs"].items()}
+            self.assertEqual(_decode(frame, confidence=0.3, outputs=arrays), _decode(frame, confidence=0.3))
+
+
+if __name__ == "__main__":
+    unittest.main()

package/python/postprocessors/test_yunet.py

@@ -0,0 +1,292 @@
+"""Pins the YuNet decoder: prior layout, box/landmark decode, landmark order.
+
+YuNet out of the box found 0 faces on 413 person crops through the SCRFD
+decoder (2026-09-26 model-replacement spike, §2.2): one prior per cell instead
+of two, centre+exp(size) boxes instead of anchor distances, and a score split
+across `cls_*` and `obj_*`. Each class below pins one of those.
+
+`unittest.TestCase`, not pytest-style: `scripts/test-python-pool.sh` runs the
+`unittest.TestCase` modules on CI's bare interpreter — the self-hosted runner
+has NO numpy. So everything that decides is plain Python and tested here
+without it: prior layout, decode, landmark order, the score, NMS, the top-K cap,
+and the output gate (`check_outputs` runs before `postprocess_yunet` touches
+numpy, so its refusals are exercised through `postprocess_yunet` itself).
+
+What genuinely needs real numpy, and is SKIPPED (named) without it: the
+vectorised pre-filter (`np.maximum(cls, obj) >= conf`), and the decode of a
+live prior through `postprocess_yunet` end to end (flatten/reshape of the
+tensors, the cap applied inside it). Those run wherever numpy is installed —
+every pool host — and in the spike harness.
+"""
+import math
+import sys
+import types
+import unittest
+
+try:  # real numpy when the sandbox has it
+    import numpy as _numpy_probe  # noqa: F401
+except ImportError:
+    _numpy_probe = None
+
+# `ndarray` is real-numpy-only: a stub another module installed would import.
+_HAS_REAL_NUMPY = _numpy_probe is not None and hasattr(_numpy_probe, "ndarray")
+
+# Without numpy, `yunet.py`'s `import numpy` is satisfied by an EMPTY stub for
+# the duration of the import only, and the stub is removed again at once: a
+# stub left in `sys.modules` would be picked up by the next test module in the
+# same `python3 -m unittest` process (test_scrfd needs `np.ceil` from its OWN
+# stub), making the result depend on module order. `yunet` keeps its own
+# reference; nothing it runs here touches it.
+if _numpy_probe is None:
+    sys.modules["numpy"] = types.ModuleType("numpy")
+    try:
+        import yunet  # noqa: F401
+    finally:
+        del sys.modules["numpy"]
+
+from yunet import (  # noqa: E402
+    NUM_LANDMARKS,
+    STRIDES,
+    TOP_K,
+    cap_top_k,
+    check_outputs,
+    decode_prior,
+    grid_cols,
+    missing_outputs,
+    nms,
+    postprocess_yunet,
+    prior_cell,
+    score_of,
+    to_source,
+    value_count,
+)
+
+
+class PriorLayoutTest(unittest.TestCase):
+    """One prior per cell, row-major — NOT SCRFD's two anchors per cell."""
+
+    def test_grid_is_one_prior_per_cell_at_640(self):
+        self.assertEqual([grid_cols(s, 640) for s in STRIDES], [80, 40, 20])
+        # 6400 + 1600 + 400. SCRFD's layout would be twice this (16800).
+        self.assertEqual(sum(grid_cols(s, 640) ** 2 for s in STRIDES), 8400)
+
+    def test_index_is_row_major(self):
+        cols = grid_cols(8, 640)
+        self.assertEqual(prior_cell(0, cols), (0, 0))
+        self.assertEqual(prior_cell(1, cols), (0, 1))  # SCRFD: index 1 is still cell (0,0)
+        self.assertEqual(prior_cell(cols + 2, cols), (1, 2))
+        self.assertEqual(prior_cell(cols * cols - 1, cols), (cols - 1, cols - 1))
+
+
+class DecodeTest(unittest.TestCase):
+    def test_box_is_centre_plus_exp_size_from_the_cell_corner(self):
+        # cell (row 1, col 2) at stride 8; centre offset (0.5, 0.5) → centre
+        # ((2+.5)*8, (1+.5)*8) = (20, 12); log size 0 → 8×8.
+        box, _ = decode_prior(8, 1, 2, [0.5, 0.5, 0.0, 0.0], [0.0] * 10)
+        self.assertEqual(box, (16.0, 8.0, 24.0, 16.0))
+
+    def test_size_is_exponential(self):
+        box, _ = decode_prior(16, 0, 0, [0.0, 0.0, math.log(3.0), math.log(2.0)], [0.0] * 10)
+        x1, y1, x2, y2 = box
+        self.assertAlmostEqual(x2 - x1, 48.0, places=4)  # 3 * 16
+        self.assertAlmostEqual(y2 - y1, 32.0, places=4)  # 2 * 16
+
+    def test_landmarks_are_cell_relative_in_stride_units(self):
+        kps = [0.25, 0.5, 0.75, 0.5, 0.5, 1.0, 0.25, 1.5, 0.75, 1.5]
+        _, points = decode_prior(32, 3, 4, [0.0, 0.0, 0.0, 0.0], kps)
+        self.assertEqual(len(points), NUM_LANDMARKS)
+        self.assertEqual(points[0], ((4 + 0.25) * 32, (3 + 0.5) * 32))
+        self.assertEqual(points[2], ((4 + 0.5) * 32, (3 + 1.0) * 32))
+
+    def test_score_is_geometric_mean_of_clamped_cls_and_obj(self):
+        self.assertAlmostEqual(score_of(0.81, 1.0), 0.9)
+        self.assertAlmostEqual(score_of(0.64, 0.25), 0.4)
+        self.assertEqual(score_of(1.3, 1.0), 1.0)
+        self.assertEqual(score_of(-0.2, 1.0), 0.0)
+
+    def test_to_source_undoes_the_letterbox(self):
+        # 320×160 crop letterboxed into 640: scale 2, pad (0, 160).
+        self.assertEqual(to_source(100.0, 260.0, 2.0, (0, 160)), (50.0, 50.0))
+
+
+class LandmarkOrderTest(unittest.TestCase):
+    """YuNet emits the SUBJECT's right eye first = image-LEFT eye, which is
+    the order `face-align.ts` maps onto ARCFACE_TEMPLATE_112. The decoder must
+    pass the five points through in model order — a sort or a swap here
+    mirrors every aligned face."""
+
+    def test_points_keep_model_order(self):
+        # A frontal face: image-left eye, image-right eye, nose, mouth L, mouth R.
+        kps = [0.2, 0.3, 0.8, 0.3, 0.5, 0.55, 0.3, 0.8, 0.7, 0.8]
+        _, points = decode_prior(8, 0, 0, [0.5, 0.5, 0.0, 0.0], kps)
+        xs = [p[0] for p in points]
+        self.assertLess(xs[0], xs[1])  # eye 0 is image-left
+        self.assertLess(xs[3], xs[4])  # mouth 3 is image-left
+        self.assertEqual([round(x, 3) for x in xs], [1.6, 6.4, 4.0, 2.4, 5.6])
+
+
+class MissingOutputsTest(unittest.TestCase):
+    """Name-only matching: a model without YuNet's twelve named tensors is
+    refused, not decoded as "no faces"."""
+
+    def test_all_twelve_present(self):
+        names = [f"{k}_{s}" for k in ("cls", "obj", "bbox", "kps") for s in STRIDES]
+        self.assertEqual(missing_outputs(names), [])
+
+    def test_names_what_is_missing(self):
+        names = [f"{k}_{s}" for k in ("cls", "bbox", "kps") for s in STRIDES]
+        self.assertEqual(missing_outputs(names), ["obj_8", "obj_16", "obj_32"])
+
+    def test_scrfd_style_outputs_are_refused(self):
+        names = [f"{k}_{s}" for k in ("score", "bbox", "kps") for s in STRIDES]
+        self.assertIn("cls_8", missing_outputs(names))
+
+
+def _list_outputs(input_size=640, drop=()):
+    """Correctly-SHAPED outputs as nested lists — no numpy needed."""
+    out = {}
+    for s in STRIDES:
+        n = grid_cols(s, input_size) ** 2
+        for kind, width in (("cls", 1), ("obj", 1), ("bbox", 4), ("kps", 10)):
+            if f"{kind}_{s}" not in drop:
+                out[f"{kind}_{s}"] = [[[0.0] * width for _ in range(n)]]
+    return out
+
+
+class OutputGateTest(unittest.TestCase):
+    """`check_outputs` runs before any numpy call, so these go through
+    `postprocess_yunet` itself on the bare interpreter."""
+
+    def test_missing_output_raises_instead_of_reporting_no_faces(self):
+        raw = _list_outputs(drop=("obj_32",))
+        with self.assertRaises(ValueError) as ctx:
+            postprocess_yunet(raw, {"confidence": 0.5}, 640, 640, 1.0, (0, 0))
+        self.assertIn("obj_32", str(ctx.exception))
+
+    def test_wrong_input_size_raises(self):
+        raw = _list_outputs(input_size=320)
+        with self.assertRaises(ValueError) as ctx:
+            postprocess_yunet(raw, {"confidence": 0.5, "inputSize": 640}, 640, 640, 1.0, (0, 0))
+        self.assertIn("does not match the model", str(ctx.exception))
+
+    def test_correct_outputs_pass_the_gate(self):
+        check_outputs(_list_outputs(), 640)  # no raise
+
+    def test_value_count_of_nested_lists(self):
+        self.assertEqual(value_count([[[0.0] * 4 for _ in range(3)]]), 12)
+        self.assertEqual(value_count(7.0), 1)
+
+
+class TopKCapTest(unittest.TestCase):
+    """The pure-Python NMS is O(n·kept): 5000 disjoint candidates took 2.8 s.
+    At most TOP_K, the best by score, may reach it."""
+
+    def _cands(self, n):
+        return [{"score": (i % 997) / 997.0, "_xyxy": (i, 0, i + 1, 1)} for i in range(n)]
+
+    def test_under_the_cap_is_untouched(self):
+        cands = self._cands(10)
+        kept, dropped = cap_top_k(cands)
+        self.assertIs(kept, cands)
+        self.assertEqual(dropped, 0)
+
+    def test_over_the_cap_keeps_the_best_k(self):
+        cands = self._cands(TOP_K + 250)
+        kept, dropped = cap_top_k(cands)
+        self.assertEqual(len(kept), TOP_K)
+        self.assertEqual(dropped, 250)
+        worst_kept = min(c["score"] for c in kept)
+        best_dropped = max(c["score"] for c in cands if c not in kept)
+        self.assertGreaterEqual(worst_kept, best_dropped)
+
+    def test_cap_is_well_below_upstream_and_above_real_frames(self):
+        # upstream 5000 (C++ NMS); measured max 110 candidates/frame at 0.1.
+        self.assertLessEqual(TOP_K, 5000)
+        self.assertGreaterEqual(TOP_K, 110 * 5)
+
+
+class NmsTest(unittest.TestCase):
+    def test_keeps_best_of_overlapping_and_disjoint(self):
+        cands = [
+            {"score": 0.6, "_xyxy": (0, 0, 10, 10)},
+            {"score": 0.9, "_xyxy": (1, 1, 11, 11)},
+            {"score": 0.7, "_xyxy": (50, 50, 60, 60)},
+        ]
+        kept = nms(cands, 0.45)
+        self.assertEqual([k["score"] for k in kept], [0.9, 0.7])
+
+
+def _outputs(np, input_size=640, live=None):
+    """Raw YuNet-shaped outputs, all priors dead except `live`:
+    {stride: (index, cls, obj, bbox4, kps10)}."""
+    out = {}
+    for s in STRIDES:
+        n = grid_cols(s, input_size) ** 2
+        out[f"cls_{s}"] = np.zeros((1, n, 1), np.float32)
+        out[f"obj_{s}"] = np.zeros((1, n, 1), np.float32)
+        out[f"bbox_{s}"] = np.zeros((1, n, 4), np.float32)
+        out[f"kps_{s}"] = np.zeros((1, n, 10), np.float32)
+    for s, (index, cls, obj, bbox, kps) in (live or {}).items():
+        out[f"cls_{s}"][0, index, 0] = cls
+        out[f"obj_{s}"][0, index, 0] = obj
+        out[f"bbox_{s}"][0, index] = bbox
+        out[f"kps_{s}"][0, index] = kps
+    return out
+
+
+@unittest.skipUnless(_HAS_REAL_NUMPY, "postprocess_yunet round trip needs real ndarray ops")
+class PostprocessYunetRoundTripTest(unittest.TestCase):
+    def setUp(self):
+        import numpy as np
+
+        self.np = np
+
+    def test_one_live_prior_decodes_to_source_pixels(self):
+        cols = grid_cols(16, 640)
+        index = 10 * cols + 20  # row 10, col 20
+        kps = [0.2, 0.3, 0.8, 0.3, 0.5, 0.55, 0.3, 0.8, 0.7, 0.8]
+        raw = _outputs(self.np, live={16: (index, 0.81, 1.0, [0.5, 0.5, math.log(4), math.log(4)], kps)})
+        # 1280×640 frame letterboxed into 640: scale 0.5, pad (0, 160).
+        out = postprocess_yunet(raw, {"confidence": 0.5, "inputSize": 640}, 1280, 640, 0.5, (0, 160))
+        dets = out["detections"]
+        self.assertEqual(len(dets), 1)
+        d = dets[0]
+        self.assertEqual(d["class"], "face")
+        self.assertAlmostEqual(d["score"], 0.9, places=3)
+        # network centre (328, 168), size 64 → [296,136,360,200] → source /0.5, y-160.
+        self.assertEqual(d["bbox"], [592.0, 0.0, 720.0, 80.0])
+        self.assertEqual(len(d["landmarks"]), NUM_LANDMARKS)
+        self.assertEqual(d["landmarks"][0], {"x": round((20.2 * 16) / 0.5, 1), "y": round((10.3 * 16 - 160) / 0.5, 1)})
+        self.assertLess(d["landmarks"][0]["x"], d["landmarks"][1]["x"])
+
+    def test_score_uses_both_tensors(self):
+        cols = grid_cols(8, 640)
+        # cls alone high, obj low → sqrt(0.9*0.1)=0.3 < 0.5: dropped.
+        raw = _outputs(self.np, live={8: (cols + 1, 0.9, 0.1, [0.5, 0.5, 1.0, 1.0], [0.0] * 10)})
+        out = postprocess_yunet(raw, {"confidence": 0.5}, 640, 640, 1.0, (0, 0))
+        self.assertEqual(out["detections"], [])
+
+    def test_prior_with_low_cls_but_high_obj_still_passes(self):
+        # sqrt(0.45 * 0.95) = 0.654 >= 0.5 although cls alone is under it: a
+        # prefilter on ONE tensor would drop a real face here.
+        cols = grid_cols(32, 640)
+        raw = _outputs(self.np, live={32: (cols + 3, 0.45, 0.95, [0.5, 0.5, 0.0, 0.0], [0.0] * 10)})
+        out = postprocess_yunet(raw, {"confidence": 0.5}, 640, 640, 1.0, (0, 0))
+        self.assertEqual(len(out["detections"]), 1)
+        self.assertAlmostEqual(out["detections"][0]["score"], 0.6538, places=3)
+
+    def test_cap_applies_inside_postprocess(self):
+        # TOP_K + 50 disjoint live priors on the stride-8 grid: all pass the
+        # score gate, none overlap, so only the cap can remove any.
+        n = grid_cols(8, 640) ** 2
+        raw = _outputs(self.np)
+        live = list(range(0, n, 2))[: TOP_K + 50]
+        raw["cls_8"][0, live, 0] = 0.9
+        raw["obj_8"][0, live, 0] = 0.9
+        raw["bbox_8"][0, live] = [0.5, 0.5, -1.0, -1.0]  # ~3 px boxes, disjoint
+        out = postprocess_yunet(raw, {"confidence": 0.5}, 640, 640, 1.0, (0, 0))
+        self.assertEqual(len(out["detections"]), TOP_K)
+
+
+if __name__ == "__main__":
+    unittest.main()

package/python/postprocessors/testdata/ssdlite_mobiledet_outputs.json

@@ -0,0 +1 @@
+{"_provenance":"Raw output tensors of google-coral/test_data ssdlite_mobiledet_coco_qat_postprocess.tflite (sha256 32c486140391eb4dc43fca7113ad392be632dc5366687f2731f73d740678693f; the CPU build of the same QAT graph as ssdlite_mobiledet_coco_qat_postprocess_edgetpu.tflite, whose four outputs have the identical names, shapes and dtypes), captured 2026-09-26 on ai-edge-litert 2.1.3 through inference_pool.py's edgetpu load/preprocess/predict path with a CPU interpreter, letterbox 320. Frames are COCO val2017 images; only the model's numeric outputs are stored here, never the pixels.","frames":{"coco_val2017_000000039769_two_cats":{"orig":[640,480],"scale":0.5,"pad":[0,40],"outputs":{"0":[[[0.15955263376235962,0.534464955329895,0.713003933429718,0.9716978073120117],[0.2043571174144745,0.02350570261478424,0.8648291826248169,0.5011452436447144],[0.10257476568222046,0.014032959938049316,0.9505413174629211,0.996590256690979],[0.22124464809894562,0.04488908499479294,0.3162550926208496,0.29261064529418945],[0.03658798336982727,0.012482762336730957,0.40508130192756653,0.987517237663269],[0.1555221676826477,0.6529091000556946,0.8028101325035095,1.006462574005127],[0.24686561524868011,0.5220372080802917,0.4193393588066101,0.5692116618156433],[0.2072891891002655,0.1583276093006134,0.3572945296764374,0.4583382308483124],[0.23659078776836395,0.5290685892105103,0.42224830389022827,0.5870484113693237],[0.5743775367736816,0.23019926249980927,0.6594175100326538,0.34779107570648193],[0.3052724301815033,0.24310186505317688,0.9099631309509277,0.9973288774490356],[0.11994627118110657,0.3792511224746704,0.5946398973464966,0.9894982576370239],[0.2084560990333557,0.2897113263607025,0.3498774766921997,0.4769533574581146],[0.048862308263778687,-0.024405688047409058,0.8030894994735718,0.6076090335845947],[0.47406071424484253,0.039728641510009766,0.8791598677635193,0.9857875108718872],[0.20506274700164795,0.006122678518295288,0.3345203399658203,0.26888036727905273],[0.21482816338539124,0.2670758366584778,0.4976709187030792,0.48709094524383545],[0.17930985987186432,0.5188078880310059,0.6031879782676697,0.87400221824646],[0.5569137930870056,0.8219212889671326,0.6033983826637268,0.91482013463974],[0.5594297051429749,0.06481415033340454,0.6947370171546936,0.3435206711292267],[0.2410658895969391,0.526638925075531,0.3487098515033722,0.5821121335029602],[0.24299399554729462,0.8754947781562805,0.3305767774581909,0.9966028332710266],[0.09360367059707642,-0.027666032314300537,0.4855625033378601,0.7005844712257385],[0.21883206069469452,0.37879708409309387,0.3296947479248047,0.46383705735206604],[0.2706087529659271,0.8736841082572937,0.34845462441444397,0.9984135031700134],[0.23773013055324554,0.5050985217094421,0.5164366364479065,0.5824022889137268],[0.8736730217933655,0.004488527774810791,1.0054889917373657,0.9902228713035583],[0.23501813411712646,0.1322900503873825,0.3002501130104065,0.2705683708190918],[0.2720530033111572,0.9462981820106506,0.35584938526153564,0.9963359236717224],[0.247519388794899,0.944722592830658,0.3231050968170166,0.9993845820426941],[0.38305434584617615,0.378726065158844,0.47873106598854065,0.49631792306900024],[0.5535222291946411,0.7801342010498047,0.7023705244064331,0.8831243515014648],[0.7894991040229797,0.03270852565765381,0.9892374873161316,0.9928076267242432],[0.20900268852710724,0.05372738838195801,0.29099857807159424,0.3962726294994354],[0.2068709284067154,0.006370965391397476,0.3372364044189453,0.0244777649641037],[0.23342448472976685,-0.01042163372039795,0.3790746331214905,0.2895890474319458],[0.22075217962265015,0.22743815183639526,0.8087112307548523,0.824122965335846],[0.5976130962371826,0.05392265319824219,0.9607176780700684,0.9731613397598267],[0.11443797498941422,0.7428056597709656,0.24389559030532837,0.9905272126197815],[0.22740966081619263,0.4848347008228302,0.4475909471511841,0.5901657938957214],[0.21618488430976868,0.5098341703414917,0.44631537795066833,0.6151652336120605],[0.8084112405776978,0.0015302002429962158,0.9273114204406738,0.9038747549057007],[-0.003454655408859253,0.01914122700691223,0.29720693826675415,0.9142001867294312],[0.23717385530471802,0.4077231287956238,0.6862777471542358,1.043309211730957],[0.40243661403656006,0.4213453233242035,0.46671468019485474,0.49919119477272034],[0.038508474826812744,0.13357895612716675,0.5688574910163879,0.814859926700592],[0.2079186886548996,0.4072537422180176,0.25239354372024536,0.45305848121643066],[0.18039143085479736,0.7189949750900269,0.346691370010376,1.0018376111984253],[0.17740964889526367,0.8574067950248718,0.3975909352302551,0.9967599511146545],[0.30818599462509155,0.000822313129901886,0.6126490831375122,0.1720949113368988],[0.06299805641174316,0.032074540853500366,0.2399788498878479,0.8108158111572266],[0.13213692605495453,0.5332436561584473,0.41890597343444824,0.9667563438415527],[0.3001849353313446,0.8360094428062439,0.5748157501220703,1.0098237991333008],[0.46447721123695374,0.7584561109542847,0.6980230808258057,0.8998774290084839],[0.22825893759727478,0.5397037863731384,0.7288434505462646,0.8317379355430603],[0.21569259464740753,0.8997492790222168,0.2564050257205963,0.9679288864135742],[0.25803816318511963,0.4330609440803528,0.3111131191253662,0.4992610216140747],[0.22732794284820557,0.3999912738800049,0.4643397331237793,0.5687592029571533],[0.23802617192268372,0.5667713284492493,0.42239078879356384,0.8623947501182556],[0.34699746966362,0.2612840533256531,0.488418847322464,0.5053830146789551],[0.5587438941001892,0.22770456969738007,0.6354513764381409,0.34881260991096497],[0.5473014116287231,0.6006264090538025,0.6144839525222778,0.6729443669319153],[0.5691072940826416,0.5631829500198364,0.6691073179244995,0.6897634267807007],[0.197118878364563,0.5749090313911438,0.5393403768539429,1.0084216594696045],[0.2407025694847107,0.4220343232154846,0.41971439123153687,0.648799479007721],[0.43011459708213806,0.42714160680770874,0.49631467461586,0.4729463458061218],[0.5331307053565979,0.7939615845680237,0.61097651720047,0.8855019211769104],[0.2184237539768219,0.7220790386199951,0.6754136085510254,1.0115272998809814],[0.23764437437057495,0.8479610085487366,0.8712865710258484,0.9984578490257263],[0.12035168707370758,0.8447889089584351,0.35048216581344604,0.9947942495346069],[0.21286840736865997,0.9534194469451904,0.26364877820014954,0.9906877279281616],[0.23123027384281158,0.04802142083644867,0.31140387058258057,0.12702254951000214],[0.5154188871383667,0.7573367357254028,0.6169029474258423,0.8059217929840088],[0.697390079498291,0.05803816393017769,0.8673416376113892,0.11111311614513397],[0.22868293523788452,0.3173567056655884,0.32868295907974243,0.45768725872039795],[0.24360230565071106,0.01714465767145157,0.5480654239654541,0.2307719588279724],[0.2897666394710541,0.5481002926826477,0.4560665786266327,0.839400589466095],[0.37094348669052124,0.7636645436286926,0.6415581107139587,0.9050858616828918],[0.46699267625808716,0.5482616424560547,0.6871740221977234,0.7246556282043457],[0.27251920104026794,0.03307163715362549,0.827480673789978,0.32943978905677795],[0.03661632537841797,0.2884189486503601,0.7459261417388916,1.0189385414123535],[0.05953405797481537,0.8303914070129395,0.2862992286682129,0.9966913461685181],[0.17441785335540771,0.2090485841035843,0.27974894642829895,0.5367846488952637],[0.18664157390594482,0.5720193386077881,0.35294151306152344,0.919648289680481],[0.19373293220996857,-0.003133825957775116,0.531268835067749,0.1656341254711151],[0.27887362241744995,0.9335945248603821,0.658628523349762,0.9955716729164124],[0.32571566104888916,0.0015931539237499237,0.6534517407417297,0.07340744137763977],[0.4985097050666809,0.058443158864974976,0.6723241209983826,0.32905781269073486],[0.5864514112472534,0.15539874136447906,0.6753339767456055,0.29165494441986084],[0.5612818598747253,0.6305587291717529,0.6754595637321472,0.719441294670105],[0.513027012348175,0.5583153367042542,0.7057222723960876,0.8958513140678406],[0.004697104915976524,-0.005696093663573265,0.040883369743824005,0.03212530165910721],[0.20103901624679565,0.33512550592422485,0.35104435682296753,0.5940406918525696],[0.20186400413513184,0.4540380537509918,0.37313663959503174,0.716795802116394],[0.22820225358009338,0.3030018210411072,0.40721407532691956,0.5136640667915344],[0.43144962191581726,0.2230268269777298,0.6185503005981445,0.4936414361000061],[0.48286062479019165,0.7618263959884644,0.6039686799049377,0.8043785095214844],[0.5385960936546326,0.6290205121040344,0.6187697052955627,0.7091941237449646],[0.4680514335632324,0.7121118903160095,0.7194494009017944,0.871222198009491],[0.6098172664642334,-0.0005137994885444641,0.7401827573776245,0.04462110251188278]]],"1":[[16.0,16.0,64.0,74.0,62.0,16.0,74.0,16.0,43.0,73.0,64.0,16.0,16.0,62.0,64.0,83.0,16.0,16.0,84.0,16.0,74.0,74.0,62.0,16.0,74.0,74.0,64.0,83.0,74.0,74.0,74.0,84.0,64.0,74.0,83.0,83.0,16.0,64.0,0.0,74.0,74.0,64.0,62.0,16.0,74.0,62.0,74.0,31.0,74.0,83.0,62.0,16.0,0.0,58.0,16.0,74.0,74.0,16.0,16.0,16.0,73.0,76.0,76.0,16.0,74.0,74.0,84.0,16.0,64.0,0.0,74.0,83.0,84.0,74.0,16.0,83.0,16.0,83.0,39.0,16.0,62.0,0.0,16.0,16.0,83.0,0.0,83.0,16.0,74.0,84.0,30.0,0.0,16.0,74.0,16.0,73.0,76.0,84.0,31.0,83.0]],"2":[[0.76953125,0.76953125,0.5234375,0.39453125,0.30078125,0.24609375,0.23828125,0.19140625,0.18359375,0.17578125,0.1640625,0.15234375,0.14453125,0.14453125,0.140625,0.1328125,0.1328125,0.125,0.11328125,0.11328125,0.109375,0.109375,0.109375,0.10546875,0.10546875,0.10546875,0.10546875,0.1015625,0.1015625,0.09765625,0.09765625,0.09765625,0.09765625,0.08984375,0.08984375,0.08984375,0.0859375,0.0859375,0.08203125,0.08203125,0.08203125,0.08203125,0.08203125,0.08203125,0.078125,0.078125,0.07421875,0.07421875,0.07421875,0.07421875,0.07421875,0.07421875,0.0703125,0.0703125,0.0703125,0.06640625,0.06640625,0.06640625,0.06640625,0.06640625,0.06640625,0.06640625,0.06640625,0.06640625,0.0625,0.0625,0.0625,0.0625,0.0625,0.05859375,0.05859375,0.05859375,0.05859375,0.05859375,0.0546875,0.0546875,0.0546875,0.0546875,0.0546875,0.0546875,0.0546875,0.05078125,0.05078125,0.05078125,0.05078125,0.05078125,0.05078125,0.05078125,0.05078125,0.05078125,0.05078125,0.046875,0.046875,0.046875,0.046875,0.046875,0.046875,0.046875,0.046875,0.046875]],"3":[100.0]}},"coco_val2017_000000252219_street":{"orig":[640,428],"scale":0.5,"pad":[0,53],"outputs":{"0":[[[0.4162486791610718,0.024550773203372955,0.7795857191085815,0.1983652412891388],[0.4350929260253906,0.5026833415031433,0.7440744638442993,0.6160659193992615],[0.4274504780769348,0.8110100030899048,0.760050356388092,0.9848244190216064],[0.5311018228530884,0.3237488269805908,0.7480632066726685,0.5054173469543457],[0.6483170986175537,0.6389470100402832,0.7928438186645508,0.7346237897872925],[0.5338170528411865,0.3173767626285553,0.6368073225021362,0.3987402021884918],[0.4912804067134857,0.9553788900375366,0.6911295056343079,1.0005137920379639],[0.30877313017845154,0.8685252070426941,0.41176334023475647,1.0109384059906006],[0.28655675053596497,0.8688049912452698,0.5342783331871033,1.0041122436523438],[0.5341765880584717,0.30358779430389404,0.6143501996994019,0.3802952766418457],[0.24573861062526703,0.535965085029602,0.3307785391807556,0.599303126335144],[0.4633682072162628,0.9736822247505188,0.6661832332611084,0.9969421029090881],[0.5260692834854126,0.38162389397621155,0.6489746570587158,0.48016151785850525],[0.7162842154502869,0.5424903035163879,0.7484474778175354,0.6030901074409485],[0.3242916762828827,0.9250457882881165,0.40213754773139954,0.9984369874000549],[0.5646662712097168,0.32681071758270264,0.6617629528045654,0.4069843292236328],[0.43745797872543335,0.8556292653083801,0.506649374961853,0.8928975462913513],[0.5453707575798035,0.9623616933822632,0.6898974776268005,0.9964772462844849],[0.3656116724014282,0.9378393292427063,0.6885550022125244,0.9913268685340881],[0.6316292881965637,0.35618144273757935,0.7316293120384216,0.47908681631088257],[0.299285888671875,0.9094834923744202,0.39638257026672363,1.0051602125167847],[0.5499159693717957,0.13722729682922363,0.5882986187934875,0.20540684461593628],[0.5260947346687317,0.3105463683605194,0.6739063858985901,0.4977883994579315],[0.5227016806602478,0.9099226593971252,0.6287714838981628,0.9592286944389343],[0.5940675735473633,0.3829779624938965,0.6574056148529053,0.46554887294769287],[0.4623275697231293,0.8369593620300293,0.4979846775531769,0.8631285429000854],[0.5491682291030884,0.47726282477378845,0.62587571144104,0.5139861106872559],[0.2056179791688919,0.4550957679748535,0.2488015741109848,0.5022701621055603],[0.21111512184143066,0.5228007435798645,0.24625077843666077,0.552243173122406],[0.5356869101524353,0.49689340591430664,0.6069472432136536,0.5310089588165283],[0.5498236417770386,0.4632295072078705,0.7031227350234985,0.511814534664154],[0.46009230613708496,0.7904802560806274,0.7690738439559937,0.9199378490447998],[0.5481249690055847,0.4862275719642639,0.7283921837806702,0.5313624739646912],[0.5189509391784668,0.02130730450153351,0.7701113224029541,0.44223350286483765],[0.6050348877906799,0.9128004312515259,0.6596968770027161,0.966663122177124],[0.4300398528575897,0.8400573134422302,0.4641554057598114,0.8629769682884216],[0.3168220520019531,0.9083377122879028,0.6748456954956055,0.9979112148284912],[0.5227682590484619,0.8477926850318909,0.6257585287094116,0.913992702960968],[0.7170714735984802,0.3222942352294922,0.7432406544685364,0.360115647315979],[0.49940651655197144,0.5581434369087219,0.5609056353569031,0.6006955504417419],[0.48628464341163635,0.40381699800491333,0.730383574962616,0.5295159816741943],[0.5414602756500244,0.5834677815437317,0.6011738777160645,0.6092543005943298],[0.7611274719238281,0.0054911356419324875,0.8006578683853149,0.03861616551876068],[0.5396264791488647,0.10698680579662323,0.760373592376709,0.5930132269859314],[0.17262639105319977,0.8333241939544678,0.4315415620803833,1.0020921230316162],[0.48002371191978455,0.8558622598648071,0.5729225873947144,0.9073963165283203],[0.5417873859405518,0.05324568599462509,0.6141053438186646,0.11295924335718155],[0.5150630474090576,0.8205909729003906,0.6275711059570312,0.8500334024429321],[0.5987178683280945,0.4551270604133606,0.7453894019126892,0.5037120580673218],[0.23898977041244507,0.5919873118400574,0.33752739429473877,0.6653785109519958],[0.2905657887458801,0.5573469996452332,0.3343902826309204,0.6073847413063049],[0.41652873158454895,0.8505421876907349,0.4688275158405304,0.8861992359161377],[0.4871012568473816,0.10838092118501663,0.5746840834617615,0.1622435450553894],[0.5822117328643799,0.850226879119873,0.6545296907424927,0.914504885673523],[0.6951959729194641,0.34702178835868835,0.737125813961029,0.4853001534938812],[0.5352706909179688,0.10564260184764862,0.6014707088470459,0.17087455093860626],[0.6321781277656555,0.9626111388206482,0.6927779316902161,0.9962278008460999],[0.7240806818008423,0.4945259094238281,0.7494900226593018,0.5657862424850464],[0.7278625965118408,0.6477597951889038,0.7912006378173828,0.699293851852417],[0.44048047065734863,0.9741995334625244,0.5787588357925415,0.9905321598052979],[0.5262652635574341,0.526103138923645,0.5561445951461792,0.550413966178894],[0.5032247304916382,0.8301358819007874,0.5570874214172363,0.8566934466362],[0.5438975691795349,0.31178784370422363,0.688424289226532,0.37798792123794556],[0.741370677947998,0.0036882534623146057,0.7968438863754272,0.1787216067314148],[0.26776328682899475,0.8388197422027588,0.4030705988407135,1.0153470039367676],[0.4859636723995209,0.8949870467185974,0.5802412629127502,0.956486165523529],[0.6024523973464966,0.49162277579307556,0.7387086153030396,0.5348063707351685],[0.5774096250534058,0.6351754665374756,0.797590970993042,0.762739896774292],[0.2301507294178009,0.5263333320617676,0.3286883533000946,0.5708081722259521],[0.2292805016040802,0.6279017925262451,0.3369244635105133,0.6867420673370361],[0.5606799125671387,0.027291707694530487,0.5996322631835938,0.05806448310613632],[0.5880016088485718,0.5785675644874573,0.6782033443450928,0.6112081408500671],[0.718255341053009,0.01484508067369461,0.7508959174156189,0.13957443833351135],[0.7315396666526794,0.8035227060317993,0.7597091794013977,0.837638258934021],[0.2944031357765198,0.809653639793396,0.453513503074646,0.9861807823181152],[0.6220482587814331,0.33179566264152527,0.7205859422683716,0.4476678669452667],[0.42942866683006287,0.7300208806991577,0.7904558181762695,1.0069555044174194],[0.44813793897628784,0.14109531044960022,0.46520859003067017,0.15604627132415771],[0.5543093681335449,0.5982905626296997,0.6266273260116577,0.6222459077835083],[0.4242232143878937,0.5523462295532227,0.7424439191818237,0.6393190622329712],[0.5669523477554321,0.5963963866233826,0.6992526054382324,0.6285596489906311],[0.21041908860206604,0.4460771381855011,0.34374770522117615,0.683089017868042],[0.5402801036834717,0.27855896949768066,0.7388873100280762,0.4006084203720093],[0.5964339971542358,0.8200300335884094,0.6535660028457642,0.8447017073631287],[0.2541125416755676,0.4373975098133087,0.3120924234390259,0.5199684500694275],[0.5770773887634277,0.04363440349698067,0.6699762344360352,0.12993639707565308],[0.7549723386764526,0.26059067249298096,0.7994471788406372,0.3144533038139343],[0.39701393246650696,0.9599695801734924,0.46035197377204895,0.9885571599006653],[0.5154039859771729,0.14906154572963715,0.5684789419174194,0.20535801351070404],[0.5914602875709534,0.08590969443321228,0.6511738896369934,0.16261713206768036],[0.6257311701774597,0.5529274940490723,0.7242688536643982,0.602965235710144],[0.7689149975776672,0.748873770236969,0.7928703427314758,0.7981798052787781],[0.212331622838974,0.4396534562110901,0.24356108903884888,0.46190768480300903],[0.5010358691215515,0.03401685133576393,0.5607494711875916,0.046919818967580795],[0.5261155366897583,0.33853837847709656,0.5592405796051025,0.390837162733078],[0.43047335743904114,0.00413740798830986,0.7486940622329712,0.07919672131538391],[0.5710752606391907,0.0036823004484176636,0.6921833157539368,0.06252261996269226],[0.6423197388648987,0.6038827300071716,0.8014300465583801,0.7752847075462341],[0.5016130208969116,0.3145033121109009,0.7622942924499512,0.5751845240592957],[0.497750461101532,0.720516562461853,0.561088502407074,0.7368491888046265]]],"1":[[0.0,0.0,0.0,61.0,32.0,26.0,0.0,27.0,27.0,26.0,0.0,0.0,26.0,40.0,27.0,26.0,0.0,30.0,0.0,61.0,27.0,30.0,26.0,30.0,26.0,0.0,26.0,0.0,0.0,30.0,61.0,0.0,61.0,14.0,30.0,0.0,0.0,30.0,36.0,30.0,0.0,30.0,40.0,14.0,27.0,26.0,30.0,30.0,61.0,71.0,0.0,0.0,26.0,30.0,61.0,30.0,32.0,40.0,40.0,0.0,30.0,30.0,26.0,40.0,27.0,30.0,61.0,32.0,0.0,61.0,0.0,61.0,40.0,36.0,27.0,61.0,0.0,76.0,30.0,0.0,61.0,71.0,0.0,30.0,83.0,30.0,36.0,27.0,42.0,30.0,40.0,40.0,0.0,0.0,26.0,0.0,14.0,32.0,14.0,0.0]],"2":[[0.76953125,0.76953125,0.76953125,0.55859375,0.46484375,0.453125,0.453125,0.39453125,0.375,0.36328125,0.3515625,0.34375,0.33203125,0.33203125,0.3203125,0.30078125,0.28125,0.28125,0.2734375,0.2734375,0.24609375,0.24609375,0.24609375,0.24609375,0.24609375,0.23828125,0.23828125,0.23046875,0.23046875,0.22265625,0.22265625,0.22265625,0.21484375,0.21484375,0.20703125,0.19921875,0.19921875,0.19921875,0.19921875,0.19140625,0.19140625,0.19140625,0.19140625,0.19140625,0.18359375,0.18359375,0.18359375,0.18359375,0.18359375,0.17578125,0.17578125,0.17578125,0.17578125,0.17578125,0.16796875,0.1640625,0.1640625,0.1640625,0.1640625,0.15625,0.15625,0.15625,0.15625,0.15625,0.15234375,0.15234375,0.15234375,0.15234375,0.14453125,0.14453125,0.14453125,0.14453125,0.14453125,0.14453125,0.140625,0.140625,0.140625,0.1328125,0.1328125,0.1328125,0.1328125,0.12890625,0.12890625,0.12890625,0.125,0.125,0.125,0.1171875,0.1171875,0.1171875,0.1171875,0.1171875,0.11328125,0.11328125,0.11328125,0.11328125,0.11328125,0.11328125,0.11328125,0.109375]],"3":[100.0]}}}}