Does π0.5 run on Unitree G1?
Capabilities
| Capability | State | Notes | Evidence |
|---|---|---|---|
| Train / fine-tune | works | Upstream fine-tunes from lerobot/pi05_base on a G1 dataset recorded with the Homunculus exoskeleton (bfloat16, gradient checkpointing, --policy.device=cuda). | ev-lerobot-g1-train, ev-unitree-il-pi05, ev-hub-xiaopengwu-pi05-g1 |
| Real-robot inference | works | Deployed with lerobot-rollout and --inference.type=rtc on --robot.is_simulation=false. Walking is handled by a separate locomotion controller chosen with --robot.controller (HolosomaLocomotionController in the documented command; GrootLocomotionController is also supported); the page does not present π0.5 as the walking controller. | ev-lerobot-g1-rollout |
| Simulation inference | no verified evidence | — | |
| Teleop + record | works | The documented recording command runs in MuJoCo (--robot.is_simulation=true) with the Homunculus exoskeleton or a joystick. | ev-lerobot-g1-intro |
Evidence 5 sources; the cell shows the highest tier
- T1 · upstream documentedofficial-docssource 2026-08-03verified 2026-10-24 (1d ago)lerobot 0.6.1
LeRobot supports the Unitree G1 (29 and 23 DoF) for teleoperation, locomanipulation policy training and simulation.
“The Unitree G1 humanoid is now supported in LeRobot! You can teleoperate, train locomanipulation policies, test in sim, and more. Both 29 and 23 DoF variants are supported.”
https://github.com/huggingface/lerobot/blob/v0.6.1/docs/source/unitree_g1.mdx - T1 · upstream documentedofficial-docssource 2026-08-03verified 2026-10-24 (1d ago)lerobot 0.6.1
The G1 page's 'Training & Inference' section fine-tunes π0.5 (--policy.type=pi05, pretrained lerobot/pi05_base).
“--policy.type=pi05 \ --output_dir=./outputs/pi05_training”
https://github.com/huggingface/lerobot/blob/v0.6.1/docs/source/unitree_g1.mdx - T1 · upstream documentedofficial-docssource 2026-08-03verified 2026-10-24 (1d ago)lerobot 0.6.1
Upstream deploys the trained policy on the physical G1 with lerobot-rollout, RTC inference and the Holosoma locomotion controller.
“--robot.type=unitree_g1 \ --robot.is_simulation=false \ --robot.controller=HolosomaLocomotionController”
https://github.com/huggingface/lerobot/blob/v0.6.1/docs/source/unitree_g1.mdx - T1 · upstream documentedvendor-docsverified 2026-10-24 (1d ago)lerobot a5b29d4
Unitree's own LeRobot-based repo documents π0.5 training on its G1 Dex3 dataset (repo pins LeRobot as a submodule at commit a5b29d4).
“--dataset.repo_id=unitreerobotics/G1_Dex3_ToastedBread_Dataset \ --policy.type=pi05”
https://github.com/unitreerobotics/unitree_IL_lerobot/blob/41c2805742de879ddab2d8d6beaeaf215f876395/README.md - T3 · community reportedhub-checkpointverified 2026-10-24 (1d ago)
config.json type=pi05; train_config.json dataset nepyope/unitree_box_move_blue_full (the example dataset linked from the upstream G1 page) whose meta/info.json robot_type=unitree_g1 (550 episodes); pretrained from lerobot/pi05_base, 3000 steps.
https://huggingface.co/xiaopeng-wu/pi05_unitree_g1
π0.5 inference speed by hardware robot-independent · inference cost only
| Compute | Runtime | Precision | Cams | p50 / p95 ms | Inference | Platform | Source |
|---|---|---|---|---|---|---|---|
| RTX 4090 | llama.cpp | fp16 | 2 @224x224 | 51.7 / — | 19.34 Hz | —conditionsboard: NVIDIA RTX 4090 · ggml_cuda_f16_f32_out: 0 (README preset for RTX 4090) Inference cost only; not task success. Full VLM + action-expert pass (ViT + LLM + 10-step action expert), LIBERO inputs (agent-view + wrist at 224x224). From the EagleVLA-Edge README RTX 4090 table (P50), PR #6 configuration. The README's naive π0.5 baseline on this GPU: 162.9 ms (not filed as a separate run: the baseline precision is not stated). | T3 · community reported |
| RTX 4090 | llama.cpp | fp16 | 2 @224x224 | 58.5 / — | 17.09 Hz | —conditionsboard: NVIDIA RTX 4090 · ggml_cuda_f16_f32_out: 0 (README preset for RTX 4090) Inference cost only; not task success. Full VLM + action-expert pass (ViT + LLM + 10-step action expert), LIBERO inputs (agent-view + wrist at 224x224). From the EagleVLA-Edge README RTX 4090 table (P50). The README's naive π0.5 baseline on this GPU: 162.9 ms (not filed as a separate run: the baseline precision is not stated). | T3 · community reported |
| Jetson AGX Thor 128GB | llama.cpp | fp16 | 2 @224x224 | 132.8 / — | 7.53 Hz | —conditionsboard: Jetson Thor (paper/README: "Jetson Thor") Inference cost only; not task success. Full VLM + action-expert pass (ViT + LLM + 10-step action expert), LIBERO inputs (agent-view + wrist at 224x224). Later configuration from the EagleVLA-Edge README (PR #6); the README notes an accuracy caveat for the two-switch GEMM preset on Thor. Naive PyTorch π0.5 on the same board per the paper: 457.9 ms (2.18 Hz) (not filed as a separate run: the baseline precision is not stated). | T3 · community reported |
| Jetson AGX Thor 128GB | llama.cpp | fp16 | 2 @224x224 | 202.8 / — | 4.93 Hz | —conditionsboard: Jetson Thor (paper/README: "Jetson Thor") Inference cost only; not task success. Full VLM + action-expert pass (ViT + LLM + 10-step action expert), LIBERO inputs (agent-view + wrist at 224x224). Paper's 'control frequency' (10.25 Hz) is 1/reaction time with VLM-skipping scheduling; not executed actions/s. Board inferred to be the 128 GB T5000 class from the paper's 130 W max-power entry for Thor (T4000 is 40–70 W). Naive PyTorch π0.5 on the same board per the paper: 457.9 ms (2.18 Hz) (not filed as a separate run: the baseline precision is not stated). | T3 · community reported |
| Jetson AGX Orin 64GB | llama.cpp | fp16 | 2 @224x224 | 271.1 / — | 3.69 Hz | power_mode 50W (paper label "Orin-50W")conditionsboard: Jetson Orin, 'Orin-50W' power mode (AGX Orin 64GB inferred, see ev-orin-board-inferred) Inference cost only; not task success. Full VLM + action-expert pass (ViT + LLM + 10-step action expert), LIBERO inputs (agent-view + wrist at 224x224). Later configuration from the EagleVLA-Edge README (time cache + cuBLASLt GEMM tuning, PR #6); not in the paper. Naive PyTorch π0.5 on the same board per the paper: 1420.8 ms (0.70 Hz) (not filed as a separate run: the baseline precision is not stated). | T3 · community reported |
| Jetson AGX Orin 64GB | llama.cpp | fp16 | 2 @224x224 | 374.5 / — | 2.67 Hz | power_mode 50W (paper label "Orin-50W")conditionsboard: Jetson Orin, 'Orin-50W' power mode (AGX Orin 64GB inferred, see ev-orin-board-inferred) Inference cost only; not task success. Full VLM + action-expert pass (ViT + LLM + 10-step action expert), LIBERO inputs (agent-view + wrist at 224x224). Paper's 'control frequency' (6.93 Hz) is 1/reaction time with confidence-based scheduling that skips the VLM; it is not executed actions/s and is not recorded as control_rate_hz. Naive PyTorch π0.5 on the same board per the paper: 1420.8 ms (0.70 Hz) (not filed as a separate run: the baseline precision is not stated). | T3 · community reported |
| Jetson Orin Nano 8GB (Super) | TensorRT | fp16 | 2 @640x480 | 438.47 / 438.93 | 2.28 Hz | jetpack 7.2.1 · power_mode MAXN_SUPER, clocks pinned with jetson_clocks (GPU 1020 MHz, CPU 1728 MHz) · tensorrt 10.16.2.10conditionsboard: Jetson Orin Nano Engineering Reference Developer Kit Super (8 GB) · l4t: R39.2.1 · os: Ubuntu 24.04.4 LTS · python: 3.12.3 · desktop: running Inference cost only; not task success. 300 s run on synthetic 640x480 observations; 10 denoising steps; latency is runtime-only over a pre-materialized observation ring. Whole-board RAM in use 6.39 GB (unified memory, desktop running). Actions were gated against the stock policy (cosine >= 0.99999) before timing. | T3 · community reported |
| Jetson Orin Nano 8GB (Super) | TensorRT | fp16 | 3 @640x480 | 643.39 / 644.92 | 1.55 Hz | jetpack 7.2.1 · power_mode MAXN_SUPER, clocks pinned with jetson_clocks (GPU 1020 MHz, CPU 1728 MHz) · tensorrt 10.16.2.10conditionsboard: Jetson Orin Nano Engineering Reference Developer Kit Super (8 GB) · l4t: R39.2.1 · os: Ubuntu 24.04.4 LTS · python: 3.12.3 · desktop: running Inference cost only; not task success. 300 s run on synthetic 640x480 observations; 10 denoising steps; latency is runtime-only over a pre-materialized observation ring. Whole-board RAM in use 6.18 GB (unified memory, desktop running). Actions were gated against the stock policy (cosine >= 0.99999) before timing. | T3 · community reported |
Embed this cell
Badge for READMEs and model cards. It always shows the current state.
[](https://caniembody.pages.dev/compat/unitree-g1/pi05)<iframe src="https://caniembody.pages.dev/embed/unitree-g1/pi05" width="380" height="132" style="border:0" title="π0.5 on Unitree G1 | CanIEmbody"></iframe>History
- 2026-10-25 fecccd1 Initial release: robot policy compatibility matrix · Danilo Znamerovszkij