MiniRT1 — <200MB Strictly-Autoregressive VLA for SO-101 Pick-and-Place
Model checkpoints for the MiniRT1 project. A visuomotor policy that generates
6-DoF joint-delta actions token-by-token (strictly autoregressive), trained by
privileged teacher-student distillation + DAgger in ManiSkill3's SO-101
SO101PlaceCube-v1 environment.
- Code + full technical report: https://github.com/charliechin424/MiniRT1
- Eval protocol:
evaluation/eval_maniskill.py, 300 episodes, seed 3000,success_at_end
Checkpoints
| File | success (300 ep) | size | description |
|---|---|---|---|
v6_dino_fp16.pt |
92% | 96 MB | project best — DINOv2-S backbone VLA (fp16) |
v6_dino_int8_static_pg64.pt |
92% | 57.6 MB | W8A8 true integer-compute (chip bit-path, per-group-64 fixed-point); zero loss |
v6_dino_w4a8.pt |
91% | 57.6 MB | 4-bit weights / 8-bit activations, near-lossless |
v4_fp16.pt |
84% | 74 MB | ResNet18 backbone, no language |
v5_lang_fp16.pt |
83% | 74 MB | +frozen-MiniLM language token (instruction-conditioned VLA) |
rl_expert_teacher.pt |
86.6% | 1.2 MB | state-based PPO teacher (privileged 49-D state; distillation source) |
task_chain_squint.npz |
— | 0.15 MB | PoE guided-decoding action-token Markov prior (Phase 8/9) |
int8/W4A8 checkpoints carry quant/group_size/w_bits/a_bits fields; the
adapter (ar_policy/policy_adapter.py) rebuilds the integer-compute module
structure automatically. See the GitHub repo for the exact loading and eval code.
ablation/ — intermediate checkpoints (reproduce the paper's tables)
The root files above are the final deliverables. ablation/ holds every
intermediate model behind the report's ablation tables:
v1_fp16.pt(1%, chunk=10 exposure-bias failure),v2_fp16.pt(45%),v3_fp16.pt(52%) — the training-recipe evolution- v4/v5 int8 variants:
{v4,v5}_int8_static_{pg64,pc}.pt,v5_int8_weightonly.pt - v6 quantization sweep:
v6_int8_static_pc.pt,v6_w4a8_clip.pt,v6_w4a4_{rtn,clip,dyn}.pt(Phase 9A low-bit study) - 9C data-scaling weak students:
v6_sub{200,800,3200}_fp16.pt(trained on 200/800/3200 episodes; the "guidance vs data size" curve)
Usage
# grab a checkpoint
hf download charliechin424/MiniRT1 v6_dino_fp16.pt --local-dir ./ckpts
# then, in the MiniRT1 repo (needs the `sim` conda env + squint env):
MINIRT1_CKPT=./ckpts/v6_dino_fp16.pt \
python evaluation/eval_maniskill.py --policy ar_policy/policy_adapter.py:make_policy --episodes 300
Citation
Built on squint (SO-101 ManiSkill3 env), LeRobot, DINOv2. Methodology: Learning by Cheating (CoRL 2019), DAgger (AISTATS 2011), RT-1/OpenVLA action tokenization.