Downloads · 30 days
4
10% of all-time downloads
Yu-Zhou-Wang/smolvla_so_101_red_foam_ball_20k
smolvla_so_101_red_foam_ball_20k is a robotics model from Yu-Zhou-Wang. 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.
Earliest saved checkpoint (step 20 000) of the same run as Yu-Zhou-Wang/smolvlaso101redfoamball (60k).
Downloads · 30 days
4
10% of all-time downloads
All-time downloads
40
Public
Parameters
450M
907 MB on disk
Likes
0
Public
Click a slice to open those files.
.safetensors907 MB · 100%
How the weights are stored.
BF16447M · 99%
From the Hugging Face model README
Earliest saved checkpoint (step 20 000) of the same run as
Yu-Zhou-Wang/smolvla_so_101_red_foam_ball (60k).
Fine-tune of lerobot/smolvla_base on
Jingyi-Z/sotac episodes 0–20
(Pick up the red foam ball and place it into the container).
Action expert only, vision encoder frozen. Tactile is not used.
| Robot | SO-101 follower, 6 joint positions |
| Cameras | training remaps top → camera1, wrist → camera2; dummy third camera (empty_cameras=1) |
| Chunk size | 50 |
| Steps | 20000 (this checkpoint) |
| Batch | 8 |
lerobot-rollout \
--strategy.type=base \
--policy.path=Yu-Zhou-Wang/smolvla_so_101_red_foam_ball_20k \
--robot.type=so101_follower \
--task="Pick up the red foam ball and place it into the container"
Camera keys must be top + wrist (or remap front → top).