quangnd58/smolvla-so101-multitask

SmolVLA SO-101 multitask

lerobot/smolvla_base finetuned on hungho77/so101-multitask (SO-101 arm, 3 tasks).

Source

• Base model: lerobot/smolvla_base
• Data: hungho77/so101-multitask (LeRobot v3.0, used as-is)
• Code: huggingface/lerobot, tag v0.6.1 (lerobot-train, docs/source/smolvla.mdx)

Details

Cameras	observation.images.top → camera1, observation.images.wrist → camera2 (rename map stored in the preprocessor)
State / action	6-D joint positions (shoulder_pan … gripper), LeRobot .pos units
Chunk size	50

Name the robot cameras top and wrist, as in the recording setup.

Tasks (use these prompts verbatim):

• Pick up the banana and place it in the bot, then close the lid
• Pick blue cube and place on red cube
• Pick all cubes and place into cup

Usage

from lerobot.policies.smolvla.modeling_smolvla import SmolVLAPolicy

policy = SmolVLAPolicy.from_pretrained("quangnd58/smolvla-so101-multitask")

On the robot, pass --policy.path=quangnd58/smolvla-so101-multitask to lerobot-record.