Upload README.md with huggingface_hub

README.md

@@ -11,7 +11,7 @@ base_model:
 ---
 # Qwen3.5-2B — ExecuTorch XNNPACK 8da4w + 8-bit embedding
 
-`qwen3_5_2b_xnnpack_8da4w_e8.pte` (1487.9 MB)
+`qwen3_5_2b_xnnpack_8da4w_e8.pte` (1483.7 MB)
 
 - **Source**: Qwen/Qwen3.5-2B
 - **License**: Apache-2.0
@@ -29,15 +29,31 @@ prompt so no answer is read through the previous one's cache:
 
 | prompt | answer | decode |
 |---|---|---|
-| capital of France? | "The capital of France is **Paris**." | 18.0 tok/s |
-| 日本の首都は? | 「日本の首都は**東京都**です。」 | 19.1 tok/s |
-| 17 times 4? | opens a `<think>` block and works through it | 19.2 tok/s |
+| capital of France? | "The capital of France is **Paris**." | 33.25 tok/s |
+| 日本の首都は? | 「日本の首都は**東京都**です。」 | 33.25 tok/s |
+| 17 times 4? | opens a `<think>` block and works through it | 33.25 tok/s |
+
+All three rates are from one run with nothing else on the machine, which matters more than
+it sounds: the same file measured 16.6 tok/s while an export was running alongside.
 
 Chat template: ChatML, bos 248045, eos [248046, 248044].
 
 Not measured on a phone. The
-[0.8B](https://huggingface.co/mlboydaisuke/Qwen3.5-0.8B-ExecuTorch) runs at 10.5 tok/s on an
-iPhone 17 Pro at 651 MB, and this file is a little over twice that size.
+[0.8B](https://huggingface.co/mlboydaisuke/Qwen3.5-0.8B-ExecuTorch) was measured at 10.5
+tok/s on an iPhone 17 Pro, but on its earlier build, before
+`use_sdpa_with_kv_cache` was turned on.
+
+## The setting that doubles it
+
+`use_sdpa_with_kv_cache: True`. Upstream's own `qwen3_5` config has it off with no reason
+given, while the equally hybrid `lfm2` config has it on. Measured on the 2B in one run:
+**8.20 tok/s off, 16.64 on**.
+
+That it is only a kernel swap and not a different model was checked three ways: three of
+four prompts came out token-identical; the first-step logits agree at **corr 0.995** with
+the same top-1; and 338 tokens deep they agree at **corr 0.990**, same top-1 and top-2. The
+fourth prompt differs where a thinking model decides how long to think, which is a near-tie
+either way.
 
 ## Running it