Update model card

README.md

@@ -85,14 +85,21 @@ them locally with `build/jepa-quantize vjepa2-ac-vitg-f16.gguf out.gguf q6_k -t
 
 ## Measured
 
-Every figure below is read from a committed artifact of jepa.cpp [`c952229`](https://github.com/aselimc/jepa.cpp/commit/c952229) by `scripts/hf_publish.py` — [parity](https://aselimc.github.io/jepa.cpp/parity/), [quantization](https://aselimc.github.io/jepa.cpp/quantization/), [accuracy](https://aselimc.github.io/jepa.cpp/accuracy/), [performance](https://aselimc.github.io/jepa.cpp/performance/) and `tests/results/*.json`.
+Every figure below is read from a committed artifact of jepa.cpp [`d53ab48`](https://github.com/aselimc/jepa.cpp/commit/d53ab48) by `scripts/hf_publish.py` — [parity](https://aselimc.github.io/jepa.cpp/parity/), [quantization](https://aselimc.github.io/jepa.cpp/quantization/), [accuracy](https://aselimc.github.io/jepa.cpp/accuracy/), [performance](https://aselimc.github.io/jepa.cpp/performance/) and `tests/results/*.json`.
+
+**Speed** — the encoder graph at f16 on 32 threads (AMD Ryzen Threadripper PRO 7995WX 96-Cores): 2228 ms per 16-frame clip; 15747 ms per 64-frame clip. The same shape on NVIDIA RTX 4500 Ada Generation: 27.8 ms. Peak RSS at f16: 3092 MiB.
 
 **Plan with f16.** A rollout compounds — step h's input is step h-1's output — so the worst predicted
 token of a 2-step rollout falls from 0.9925 at f16 to 0.9368 at q8_0 to 0.5429 at q4_k, and at q4_k on
 a GPU the planning energy misranks the candidates and the model picks a different action. The encoder
 half of the bundle passes at every tier on both backends. At f32 on the CPU the predictor is exact to
 cosine 1.0000000 against Meta's own world-model dump, and K candidates batched on the graph's batch
-axis are bit-identical to K sequential rollouts. Full tables in [parity](https://aselimc.github.io/jepa.cpp/parity/).
+axis are bit-identical to K sequential rollouts, and `jepa_ac_plan` reproduces Meta's own CEM
+planner to 3e-08 on its own recorded random draws. Full tables in [parity](https://aselimc.github.io/jepa.cpp/parity/).
+
+**`-q4_k.gguf` self-reports `general.file_type = q4_0`, and that is correct.** K-quants need
+`ne[0] % 256 == 0` and this family's width is 1408 = 5.5 x 256, so every `[*, 1408]`-row matrix
+falls back to q4_0; the field is the most common stored type. `jepa-info` lists the real mix.
 
 ## Source, licence and attribution
 
@@ -110,7 +117,7 @@ The licence travels inside every GGUF as `general.license` and the origin as `ge
 
 ## Conversion
 
-Produced by jepa.cpp [`c952229`](https://github.com/aselimc/jepa.cpp/commit/c952229):
+Produced by jepa.cpp [`d53ab48`](https://github.com/aselimc/jepa.cpp/commit/d53ab48):
 
 ```bash
 scripts/download_models.sh --convert vjepa2-ac