Downloads · 30 days
13
37% of all-time downloads
alexhegit/so101-simstudio-lab01-pnp-molmoact2
so101-simstudio-lab01-pnp-molmoact2 is a robotics model from alexhegit. Use it for the robotics task on the model card, and read the license before you ship it in a product. It is set up for lerobot. The card lists the license as apache-2.0.
MolmoAct2 policy fine-tuned from lerobot/MolmoAct2-SO100101-LeRobot on expert demonstrations collected and validated with SO-101 SimStudio (MuJoCo sim2sim, leader-arm teleop).
Downloads · 30 days
13
37% of all-time downloads
All-time downloads
35
Public
Parameters
5.6B
11.5 GB on disk
Likes
0
Public
Click a slice to open those files.
.safetensors11.5 GB · 100%
How the weights are stored.
BF165.4B · 97%
From the Hugging Face model README
MolmoAct2 policy fine-tuned from lerobot/MolmoAct2-SO100_101-LeRobot on expert demonstrations collected and validated with SO-101 SimStudio (MuJoCo sim2sim, leader-arm teleop).
Training data: alexhegit/so101-simstudio-lab01-pnp
This Hub revision is the MI300X run: batch 32, 10 000 steps (~320K sample updates, ~15 epochs), VLM LoRA + trainable action expert, final train loss 0.009, ~30 GB of 192 GB HBM. Wall time ~13 h 40 m on AMD Instinct MI300X (DORobot). Checkpoint 010000 / last.
Cameras used for this fine-tune (2-view warm start matching the SO100_101-LeRobot base): camera_top → cam0, camera_wrist → cam1. Lab01 joints are radians; joint frame transform is identity (no SO100 degree offsets).
Closed-loop MuJoCo success rates are not published for this checkpoint yet.
| Resource | Link |
|---|---|
| SimStudio repo | rocPAI-Forge/so101-simstudio |
| Lab 01 walkthrough | labs/lab01_pnp/lab01_pnp.md |
from lerobot.policies.molmoact2.modeling_molmoact2 import MolmoAct2Policy
policy = MolmoAct2Policy.from_pretrained("alexhegit/so101-simstudio-lab01-pnp-molmoact2")