Upload CLAUDE.md with huggingface_hub

CLAUDE.md

@@ -6,8 +6,10 @@ Small vision-language-action models. Two lines of work:
   0.5–2 Hz) + fast flow-matching expert (10–20 Hz). **86% LIBERO-spatial, 84% 4-suite
   average** — beats OpenVLA-7B on every suite at 0.64B params / 140M online.
 - **v2 (research)** — cross-embodiment: canonical action space + morphology
-  conditioning across 10 robots. Best: `tv2_C_scaled` (ratio-to-baseline **0.81**
-  in-training / **0.88** held-out; was 0.99/1.17 before scaling).
+  conditioning across 10 robots. Best: `tv2_C_scaled` — ratio-to-baseline **0.34**
+  in-training, **0.28** on a held-out *arm*, but **1.67** on a held-out *mobile
+  base*. (Published as 0.81/0.88 until an eval-harness bug was found on
+  2026-08-28 — see Results.)
 
 Everything (checkpoints, results, wiring) lives in the **private** HF repo
 `AlexWortega/tinyvla` (v1 at the root, v2 under `tinyvla2/`).