Card: image timings on an idle NPU, 31k row

MANIFEST.sha256

@@ -4,7 +4,7 @@
 c71d239df91726fc519c6eb72d318ec65820627232b2f796219e87dcf35d0ab4  LICENSE
 4527ccda731d2c1149d8529d46ed6ce4035f19041951ba87b579d9dfe867bb33  model.q4nx
 b926ac0e9a72a24e29e3f518f3241064416bc137c185c3349c3eae2370cfe7a0  NOTICE
-1673bed3ee81b7121142e744896754325c413a6b67dcc2ab03707f7738c1760b  README.md
+b1d7211158ed046b23f571e34b9ad9239b90291a313a8714a817960c594a6074  README.md
 662e8a26d8016b2d479f1ca66bf64944603316c2b5e074ab7ac3304b8ae045e9  readout.safetensors
 49e2b6e395f959f077f1e992b338919c0d4a9732fc6e613995e06557f843500c  tokenizer_config.json
 5f9e4d4901a92b997e463c1f46055088b6cca5ca61a6522d1b9f64c4bb81cb42  tokenizer.json

README.md

@@ -60,16 +60,18 @@ suite; the Decision Index public vision suite has not been run for this package.
 | 1 question, 1,024 tokens | 584 ms |
 | 1 question, 4,096 tokens | 2,470 ms |
 | 1 question, 15,854 tokens | 11.4 s |
+| 1 question, 31,491 tokens | 27.1 s |
 | 1 question, 32,768 tokens | 28.4 s |
 | 4 questions, 330 tokens each | 796 ms |
 | 16 questions, 330 tokens each | 2,889 ms |
-| 1 question + a 1024 x 768 image (876 tokens, HTTP) | 2.2 s |
-| 1 question + a 2000 x 1500 image (1,672 tokens, HTTP) | 4.8 s |
-| 8 questions + one such image (one image encoding, HTTP) | 8.1 s |
+| 1 question + a 1024 x 768 image (876 tokens, HTTP) | 1.3 s |
+| 1 question + two images (HTTP) | 2.4 s |
+| 1 question + a 2000 x 1500 image (downscaled to 1.6 MP, HTTP) | 3.0 s |
+| 8 questions + one image (one image encoding, HTTP) | 5.0 s |
 
-From 1,536 tokens attention runs on the NPU's attention kernel (below that the CPU is faster). Prompts over 32,768
+From 1,536 tokens attention runs on the NPU's attention kernel (below that the CPU is faster; at 31,491 tokens it is 3.3x faster than attention on the CPU). Prompts over 32,768
 tokens are answered with `max_length_exceeded` (nothing is truncated). Text rows were measured at the engine (HTTP
-adds about a millisecond); the image rows over HTTP.
+adds about a millisecond); the image rows over HTTP, one question, median of 3, on an otherwise idle NPU.
 
 ## License