Add measured Galaxy S26 NPU/GPU section

Numbers from the S26 classic sweep. Primary logs:
litertlm-convert/community_accel_work/s2_npu_sweep/results.jsonl (on-device
JIT) and portal_work/bench_all.txt + bench_redo.txt (the 2026-08-23 AOT
sweep); every row states which of the two produced it. Rows carry device,
SoC, runtime, N and thermal state, and join to this repo by exact .tflite
file name. Rows taken outside the sweep's thermal window are not reported.

README.md

@@ -178,7 +178,22 @@ Android figures use the standard TFLite [`benchmark_model`](https://ai.google.de
 | `encode_512` | 1287 ms | **581 ms** |
 | `mlm_128` | 431 ms | **135 ms** |
 
-**GPU works as of the 2026-08-13 re-export — via the single-signature file.** This re-export respells the one idiom mobile GPU delegates refuse — transformers' rank-5 `repeat_kv` expand — into an equivalent rank-4 matmul (outputs bitwise-identical on CPU), which makes the graph fully GPU-delegable. The 5-signature file still cannot compile on phone GPUs: each signature subgraph materializes its own fp32 weight copy (~7 GB total for this model), which exceeds phone memory on both iOS and Android. The new **`LFM2.5-Encoder-350M_wi8fc_single-sig.tflite`** (`encode_512` only, 359 MB) is the phone-GPU artifact: Pixel 8a OpenCL full delegation, `encode_512` **12.2 ms**, cosine **0.9994** vs the fp32 desktop reference (LiteRT CompiledModel API, fp32 GPU precision, best of 3 warm runs); iPhone 17 Pro Metal 171 ms, cosine 0.9994. The 5-signature file remains the CPU artifact (bit-exact on device; iPhone `encode_512` 144–158 ms). Set the GPU precision to **fp32** — at fp16 GPU precision this family's norm reductions overflow and every output is NaN. These CompiledModel timings are not comparable to the classic-delegate `benchmark_model` timings above (different GPU runtime).
+**GPU works as of the 2026-08-13 re-export — via the single-signature file.** This re-export respells the one idiom mobile GPU delegates refuse — transformers' rank-5 `repeat_kv` expand — into an equivalent rank-4 matmul (outputs bitwise-identical on CPU), which makes the graph fully GPU-delegable. The 5-signature file does not compile on the Pixel 8a OpenCL or iPhone 17 Pro Metal delegates: each signature subgraph materializes its own fp32 weight copy (~7 GB total for this model), which exceeds those phones' memory. It does compile and run on the Adreno GPU of a Galaxy S26 under LiteRT `CompiledModel` 2.2.0 — 82.04 ms — but it pays a 17.3 s load there against 4.07 s for the single-signature file (see Snapdragon NPU (Hexagon) below). The new **`LFM2.5-Encoder-350M_wi8fc_single-sig.tflite`** (`encode_512` only, 359 MB) is the phone-GPU artifact: Pixel 8a OpenCL full delegation, `encode_512` **12.2 ms**, cosine **0.9994** vs the fp32 desktop reference (LiteRT CompiledModel API, fp32 GPU precision, best of 3 warm runs); iPhone 17 Pro Metal 171 ms, cosine 0.9994. The 5-signature file remains the CPU artifact (bit-exact on device; iPhone `encode_512` 144–158 ms). Set the GPU precision to **fp32** — at fp16 GPU precision this family's norm reductions overflow and every output is NaN. These CompiledModel timings are not comparable to the classic-delegate `benchmark_model` timings above (different GPU runtime).
+
+## Snapdragon NPU (Hexagon)
+
+- `LFM2.5-Encoder-350M_fp16.tflite` — neither accelerator produced a usable row on the S26. Both ended the same way: `LiteRtException: Failed to compile model`.
+- `LFM2.5-Encoder-350M_wi8fc.tflite` — the GPU runs it at 82.04 ms. The NPU does not — `LiteRtException: Failed to compile model`.
+- `LFM2.5-Encoder-350M_wi8fc_single-sig.tflite` — the GPU runs it at 82.24 ms. The NPU does not — the graph compiles and then fails to run (`LiteRtException: Failed to invoke the compiled model`).
+
+| file | backend | compiled | inference (median / min) | load |
+|---|---|---|---:|---:|
+| `LFM2.5-Encoder-350M_wi8fc.tflite` | GPU (Adreno) | — | 82.04 ms / 81.14 ms | 17349 ms |
+| `LFM2.5-Encoder-350M_wi8fc_single-sig.tflite` | GPU (Adreno) | — | 82.24 ms / 80.41 ms | 4074 ms |
+
+Measured on a **Samsung Galaxy S26** (Snapdragon 8 Elite Gen 5 / SM8850, Hexagon v81, Android 16) with LiteRT `CompiledModel` 2.2.0, one accelerator per process, 5 warm-up runs then N=50 timed runs, median reported. Every run held thermal status `NONE` throughout. Headroom 0.72–0.78, where 1.0 is the throttling threshold.
+
+GPU wiring: [GPU guide](https://github.com/john-rocky/hf-to-litertlm/blob/main/docs/android-gpu.md). NPU recipe: [NPU guide](https://github.com/john-rocky/hf-to-litertlm/blob/main/docs/android-npu.md).
 
 ## License