privacy: scrub host chip/OS from model card
Browse files
README.md
CHANGED
|
@@ -77,7 +77,7 @@ pip install coreai-catalog && coreai-catalog install act-aloha-cube
|
|
| 77 |
|
| 78 |
- **Gate A (structure): passed** — the bundle's layout + metadata were
|
| 79 |
validated on real hardware; the graphs load.
|
| 80 |
-
- **Gate B — action_parity: 100.0% min chunk-cosine · 0.0 max normalized MAE** vs the fp16 reference over 8 frames of `8 fixed-seed observations (deterministic ACT: the export is exact regardless of input; realistic-episode eval is future work)`, 1-step, fixed-noise (measured on Apple
|
| 81 |
- This certifies the export is **numerically faithful to the source policy** — it
|
| 82 |
does **NOT** certify real-world task success, embodiment transfer, or
|
| 83 |
closed-loop stability. Closed-loop success-rate eval (LIBERO/ManiSkill):
|
|
|
|
| 77 |
|
| 78 |
- **Gate A (structure): passed** — the bundle's layout + metadata were
|
| 79 |
validated on real hardware; the graphs load.
|
| 80 |
+
- **Gate B — action_parity: 100.0% min chunk-cosine · 0.0 max normalized MAE** vs the fp16 reference over 8 frames of `8 fixed-seed observations (deterministic ACT: the export is exact regardless of input; realistic-episode eval is future work)`, 1-step, fixed-noise (measured on Apple Silicon).
|
| 81 |
- This certifies the export is **numerically faithful to the source policy** — it
|
| 82 |
does **NOT** certify real-world task success, embodiment transfer, or
|
| 83 |
closed-loop stability. Closed-loop success-rate eval (LIBERO/ManiSkill):
|