Downloads · 30 days
16
43% of all-time downloads
pi05_roarm_m3_candy_pick_place
polrolnik2/pi05_roarm_m3_candy_pick_place
pi05_roarm_m3_candy_pick_place is a robotics model from polrolnik2. 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.
All-time downloads
37
Public
Parameters
4.1B
9.4 GB on disk
Likes
0
Public
Repo makeup
Click a slice to open those files.
.safetensors9.4 GB · 100%
Parameter types
How the weights are stored.
BF163.6B · 87%
Model card
From the Hugging Face model README
pi0.5 finetuned on rtrtwrw/roarm_m3_candy_pick_place
Trained with LeRobot 0.6.2.
| dataset (as trained) | rtrtwrw/roarm_m3_candy_pick_place |
| base checkpoint | lerobot/pi05_base |
| steps | 30000 |
| effective batch size | 64 (4 per GPU x 4 GPUs x 4 accumulation) |
| policy dtype | bfloat16 |
| seed | 1000 |
| checkpoint | last |
Input and output features
| feature | type | shape |
|---|---|---|
observation.images.base_0_rgb | VISUAL | [3, 224, 224] |
observation.images.left_wrist_0_rgb | VISUAL | [3, 224, 224] |
observation.images.right_wrist_0_rgb | VISUAL | [3, 224, 224] |
observation.state | STATE | [32] |
action | ACTION | [6] |
Camera key rename map
This dataset's camera names differ from the base model's. Training used this rename map and inference needs the same:
{
"observation.images.front": "observation.images.base_0_rgb",
"observation.images.wrist": "observation.images.left_wrist_0_rgb"
}
Usage
from lerobot.policies.factory import get_policy_class
policy = get_policy_class("pi05").from_pretrained("polrolnik2/pi05_roarm_m3_candy_pick_place")
- Task
- Robotics
- Library
- lerobot
- License
- apache-2.0
- Created
- Aug 27, 2026
- Updated
- Oct 17, 2026