Model card: simpler inputs wording

README.md

@@ -39,15 +39,9 @@ alongside actions. A single checkpoint runs on any subset of its streams, so the
 compute/accuracy operating point is a runtime flag.
 
 This checkpoint is a large-scale pretrain for bimanual 6-DoF YAM-family robots
-with one head camera and two wrist cameras.
-
-- **It always uses RGB and DINO.** The client sends RGB; the model computes the
-  DINOv3 features from it.
-- **Depth is optional.** With depth and intrinsics the 3D pointmap stream is on as
-  well; without them the model runs on RGB and DINO alone.
-- **Training depth** was DA3 (Depth Anything 3) estimates for ABC-130k, MolmoAct2
-  and AgiBot World, and the YAM recordings' own sensor depth.
-
+with one head camera and two wrist cameras. It takes RGB, with optional depth,
+and always uses DINO, which the model computes from the RGB. Its training depth
+is DA3 for ABC-130k, MolmoAct2 and AgiBot World, and sensor depth for the YAM data.
 It is a starting point for fine-tuning; zero-shot use is possible on a matching
 rig. The sibling without depth is
 [`flex-pi/flexpi-pretrain-rgb-preview`](https://huggingface.co/flex-pi/flexpi-pretrain-rgb-preview).
@@ -58,14 +52,14 @@ rig. The sibling without depth is
 |---|---|
 | Architecture | Mixture-of-Transformers: Wan2.2-TI2V-5B video expert (5.00 B) + ActionDiT action expert (1.02 B), 30 layers, HBridge 7 / 16 / 7 |
 | Cameras | `cam_high` 256×320, `cam_left_wrist` and `cam_right_wrist` 224×224, composited at 384×320 (`robotwin_uniform14`), 33 frames |
-| Inputs | three RGB cameras, plus optional metric depth and intrinsics; DINOv3 features are always used, computed by the model from the RGB frames |
+| Inputs | three RGB cameras, optional depth; DINO is computed from the RGB |
 | Visual streams | RGB and 3D pointmap through the frozen Wan2.2 VAE; DINOv3 ViT-B/16 (frozen), temporal stride 1 |
 | Flex regime | `p_present_* = p_j* = 0.5` for video, DINO and pointmap; cross-modal forcing on |
 | Action head | 34-D, shared by 6- and 7-joint arms. Channels 26 and 33 are each arm's 7th joint, which a YAM arm does not have: they are left out of the loss on YAM data and held at 0 when serving a YAM robot, which sends and receives 32-D |
 | Wire format | 32-D YAM end-effector state in, 32 × 32 absolute action chunk out |
 | Normalization | frame-pooled q01/q99 over 34 channels (`dataset_stats.json`), normalized targets unclamped (`action_norm_clamp: inf`) |
 | Training data | ~4,600 hours of robot data: ABC-130k, MolmoAct2, AgiBot World and in-house YAM teleoperation data |
-| Training depth | DA3 (Depth Anything 3) estimates for ABC-130k, MolmoAct2 and AgiBot World; the YAM recordings' own sensor depth |
+| Training depth | DA3 for ABC-130k, MolmoAct2 and AgiBot World; sensor depth for YAM |
 | Checkpoint | `step_190000`, bf16, 12,082,212,409 bytes, sha256 `b168eb68f255b98bf4805ad762dd870e06dd5f0f6f044f3e4280622338799eb2` |
 
 ## Usage