model card

README.md

@@ -5,21 +5,21 @@ pipeline_tag: robotics
 tags: [robotics, vla, gr00t, gr00t-n1.6, manipulation, maniguard, franka]
 ---
 
-# GR00T-N1.6 - Clutter (joint, 2-cam)
+# GR00T-N1.6 - Sim2Real Clutter (task_0048) (joint, 2-cam)
 
-NVIDIA Isaac **GR00T-N1.6-3B** fine-tuned on the ManiGuard **clutter** base task (sim
-Franka Panda). Part of the ManiGuard VLA benchmark - GR00T vs pi0.5 on the same task
-families with identical data, cameras, and controller.
+NVIDIA Isaac **GR00T-N1.6-3B** fine-tuned on the ManiGuard **clutter-task0048** base task (sim Franka Panda).
+Part of the ManiGuard VLA benchmark - GR00T vs pi0.5 on the same task families with
+identical data, cameras, and controller.
 
 ## Model
 - **Base:** [nvidia/GR00T-N1.6-3B](https://huggingface.co/nvidia/GR00T-N1.6-3B) - Eagle (nvidia/Eagle-Block2A-2B-v2) VLM + flow-matching DiT action head
 - **Embodiment:** NEW_EMBODIMENT - Franka Panda, **8-D joint** state/action (7 arm joints + 1 gripper)
 - **Cameras (2):** image_left (overview) + wrist (256x256)
-- **Action:** arm = state-relative chunks, gripper = absolute; 16-step horizon; NON_EEF (joint space)
+- **Action:** arm = state-relative chunks (reconstructed to absolute at inference), gripper = absolute; 16-step horizon at 30 fps (0.53 s); NON_EEF (joint space)
 - **Tuning:** GR00T-N1.6 default - VLM (LLM + visual) **frozen**, train projector + diffusion action head (**no LoRA**)
 
 ## Training
-- 8-card config, DeepSpeed ZeRO-2, bf16, global batch 4, 50000 steps (~8 epochs over 23,904 frames), cosine LR (peak 2e-4, sqrt-scaled), warmup 0.05
+- single-card config, bf16, global batch 4, 50000 steps (~8 epochs over 23,904 frames), cosine LR (peak 2.5e-5, sqrt-scaled), warmup 0.05
 - Data: [IDEAS-Lab-Northwestern/sim2real-clutter-task0048-sim-joint-5cam](https://huggingface.co/datasets/IDEAS-Lab-Northwestern/sim2real-clutter-task0048-sim-joint-5cam); videos decoded as H.264 for GR00T's torchcodec loader
 
 ## Usage