Downloads · 30 days
0
pi05-sort-three-vlm-lora-checkpoints
alanoob/pi05-sort-three-vlm-lora-checkpoints
pi05-sort-three-vlm-lora-checkpoints is a robotics model from alanoob. 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 openpi.
Public JAX checkpoints from action-only fine-tuning of PI0.5 on kitalr/pandasortthreecontinuedeefullresfullv21.
Access
Public
Updated Sep 1, 2026
Repo size
43 GB
Likes
1
Public
Repo makeup
Click a slice to open those files.
Other43 GB · 100%
At a glance
- Task
- Robotics
- Library
- openpi
- Access
- Public
- Created
- Sep 1, 2026
- Updated
- Sep 1, 2026
- SHA
- 4f84d1a3
Model card
From the Hugging Face model README
PI0.5 Sort-Three VLM-LoRA Checkpoints
Public JAX checkpoints from action-only fine-tuning of PI0.5 on
kitalr/panda_sort_three_continued_ee_fullres_full_v21.
The PaliGemma VLM uses LoRA and the action expert is fully trained. Checkpoints
are provided at steps 2k, 5k, 10k, 15k, 20k, and 30k. Each step directory
contains params, assets, and _CHECKPOINT_METADATA. Optimizer/train state is
not included.
OpenPI config name: sort_three_vlm_lora_full_action_bs16.
These checkpoints are research artifacts. The diagnostic text generated through the vocabulary decoder is not an official PI0.5 subtask-head prediction.
- Task
- Robotics
- Library
- openpi
- Created
- Sep 1, 2026
- Updated
- Sep 1, 2026