armteam/hapticwam-teacher

HapticWAM teacher

HapticWAM: Distilling Imagined Touch into a World-Action Model without Inference-Time Tactile Sensing — paper arXiv:2609.23888, submitted to ICRA 2027. Code: github.com/Advanced-Robotic-Manipulation/HapticWAM · all repos: HapticWAM — ICRA 2027.

The tactile world-action-model teacher: the full-input model that sees the fingertip tactile streams and is distilled into the deployed student.

The checkpoint the paper deploys is teacher_v6_simft/teacher_002000.pt.

Folders

Folder	Files	Size	What it is
teacher_v6_simft/	1	0.39 GB	the deployed teacher. teacher_002000.pt, v6 fine-tuned on simulation plus rig rollouts, 2,000 steps.
teacher_v6/	10	0.79 GB	the v6 teacher before the sim fine-tune: teacher_019000.pt, teacher_020000.pt, the training log train_v6.log, and eval/ with seven evaluation JSONs (steps 18k/19k/20k at NFE 1 and 5, plus a v5.6 comparison). teacher_020000.pt is the arm that runs on the rig as the teacher-only condition.
eval/	4	—	teacher evaluations at NFE 1 and 5 (v6_teacher_nfe1, v6_teacher_nfe5), each a JSON plus its log.
text_embeddings.pt	1	0.82 GB	the Cosmos text-embedding cache for the task prompts. Not a checkpoint: it is a precomputed input that the provisioning scripts download next to the teacher and place at data/phantom-episodes/tasks/. It is regenerable from the prompts, and it is kept here so that provisioning reads one repo rather than two.

index.jsonl lists every file here with its size, LFS sha256 and the path it was copied from.

Training data

Trained on armteam/hapticwam-teleop-dataset, 1,115 teleoperation episodes indexed by manifests_v6.tar. The sim fine-tune additionally uses simulation episodes and the rig rollouts in armteam/hapticwam-rollouts.

Part of the HapticWAM release

Ten repos on the hub, gathered in the HapticWAM — ICRA 2027 collection.

Repo	Kind	Holds
armteam/hapticwam-teacher ← you are here	model	the tactile-input teacher. Deployed checkpoint teacher_v6_simft/teacher_002000.pt; also holds the Cosmos prompt cache text_embeddings.pt
armteam/hapticwam-student	model	the distilled pad-free student, the model that runs on the rig. Deployed checkpoint hid_simft/student_001000.pt
armteam/hapticwam-baselines	model	the pi0.5, Diffusion Policy and X-VLA baselines at the deployed steps
armteam/hapticwam-ablations	model	every training arm that is not deployed, and the complete evaluation sweeps
armteam/hapticwam-teleop-dataset	dataset	the training corpus — 1,115 teleoperated episodes, packed per task
armteam/hapticwam-teleop-raw	dataset	the same teleoperation as loose, as-recorded sessions (provenance)
armteam/hapticwam-sim-episodes	dataset	Isaac Sim expert episodes, used for the sim fine-tune
armteam/hapticwam-rig-episodes	dataset	the closed-loop rig takes the reported numbers are computed from
armteam/hapticwam-rollouts	dataset	policy-driven rollouts — the DAgger rounds and the deploy days
armteam/hapticwam-evidence	dataset	per-take evidence behind the paper's tables — scored CSVs, probe JSONs, figures

Code, training and deployment scripts: github.com/Advanced-Robotic-Manipulation/HapticWAM.

Licence

Model weights and the HapticWAM code: Apache-2.0. The datasets listed above: CC-BY-4.0.