Upload README.md with huggingface_hub

README.md

@@ -22,13 +22,23 @@ on the next call, with the same weights.
    -> scores (1,900,256), boxes (1,900,4)
 ```
 
-| build | file | MB | scores corr | boxes corr | Mac ms* |
-|---|---|---|---|---|---|
-| fp32 | `grounding_dino_tiny_xnnpack_fp32.pte` | 690.3 | 0.999833 | 0.998992 | 584.1 |
+| build | file | MB | scores corr | boxes corr | detections agreeing | Mac ms* |
+|---|---|---|---|---|---|---|
+| fp32 | `grounding_dino_tiny_xnnpack_fp32.pte` | 690.3 | 0.999833 | 0.998992 | **26/26** | 584.1 |
+| int8 (dynamic) | `grounding_dino_tiny_xnnpack_int8.pte` | **254.4** | 0.804530 | 0.777331 | **26/27** | 738.7 |
 
 \*Mac arm64, single process, median of 5 — a reference point for relative cost, not a
-device number. Torch eager fp32 on the same machine: 376.7 ms. XNNPACK delegate
-coverage 63.0% (3503/5559 ops).
+device number. Torch eager fp32 on the same machine: 376.7 ms. XNNPACK delegate coverage
+63.0% (3503/5559 ops) for fp32, 71.5% for int8.
+
+**Read that int8 row carefully.** Its correlations are 0.80 and 0.78, which on any other
+model on this shelf would withdraw the build. Here they are dominated by the queries that
+detect nothing, whose coordinates are free to move without changing an answer — and the
+detections themselves hold: all 26 that eager finds are found, at IoU > 0.5 with the same
+prompt token, plus one box eager does not have. It is 37% of the fp32 file and it is what
+you would put on a phone, but the evidence behind it is 26 detections over five
+photographs, not a COCO run, and one of those 27 boxes already disagrees. It is also
+slower than fp32 on this Mac (738.7 ms against 584.1).
 
 **There is no Core ML build.** The graph reshapes to rank 6 and Core ML's ceiling is 5 —
 the same wall RAFT-small hits with its correlation volume. This one is XNNPACK only, and
@@ -69,11 +79,14 @@ as the label.
 The number that decides is detections, not correlation. Against eager PyTorch on five
 street photographs at threshold 0.3:
 
-**26 of 26 of eager's detections are reproduced**, each at IoU > 0.5 with the same prompt
-token. That includes a photo with 17 detections in it.
+**fp32 reproduces 26 of 26**, each at IoU > 0.5 with the same prompt token, and invents
+nothing. That includes a photo with 17 detections in it. **int8 scores 26 of 27** — it
+finds all of eager's and one box besides. The denominator is whichever build found more,
+so an invented detection costs the same as a missed one; counting only the reference's
+detections would give a build that returns every box in the image a perfect score.
 
 ```bash
-python convert/audit_int8.py grounding_dino_tiny --variant fp32
+python convert/audit_int8.py grounding_dino_tiny --variant fp32   # or int8
 python convert/verify_cards.py grounding_dino_tiny
 ```