Zum Inhalt springen
ModelÜberprüftExtern

Pi05 So101 Pick Block In Cup 2cam Run1 v2

ArtOtherVerleger@chasedreaminfCBewertung

Modellquelle

Art
Other
Dateigröße
7,36 KB–8,71 GB
Aktualisiert
13. Sept. 2026

Quellenauszug

Modellartefakte

2 Artefakte

model.safetensors

safetensors · 8,71 GB · SHA-256 1295625c7ec8…1b15

policy_preprocessor_step_2_normalizer_processor.safetensors

safetensors · 7,36 KB · SHA-256 ffa3bb37edf8…5c6d

Quellenauszüge

3 Auszüge

--- datasets: chasedreaminf/record-pick-block-in-cup-2cam-run1_20260505_191315 library_name: lerobot license: apache-2.0 model_name: pi05 pipeline_tag: robotics tags: - lerobot - robotics - pi05 --- # Model Card for pi05 <!-- Provide a quick summary of what the model is/does. --> **π₀.₅ (Pi05) Policy** π₀.₅ is a Vision-Language-Action model with open-world generalization, from Physical Intelligence. The LeRobot implementation is adapted from their open source OpenPI repository. **Model Overview** π₀.₅ represents a significant evolution from π₀, developed by Physical Intelligence to address a big challenge in robotics: open-world generalization. While robots can perform impressive tasks in controlled environments, π₀.₅ is designed to generalize to entirely new environments and situations that were never seen during training. For more details, see the [Physical Intelligence π₀.₅ blog post](https://www.physicalintelligence.company/blog/pi05). This policy has been trained and pushed to the Hub using [LeRobot](https://github.com/huggingface/lerobot). See the full documentation at [LeRobot Docs](https://huggingface.co/docs/lerobot/index). --- ## How to Get Started with the Model For a complete walkthrough, see the [training guide](https://huggingface.co/docs/lerobot/il_robots#train-a-policy). Below is the short version on how to train and run inference/eval: ### Train from scratch ```bash lerobot-train \ --dataset.repo_id=${HF_USER}/<dataset> \ --policy.type=act \ --output_dir=outputs/train/<desired_policy_repo_id> \ --job_name=lerobot_training \ --policy.device=cuda \ --policy.repo_id=${HF_USER}/<desired_policy_repo_id> --wandb.enable=true ``` _Writes checkpoints to `outputs/train/<desired_policy_repo_id>/checkpoints/`._ ### Evaluate the policy/run inference ```bash lerobot-record \ --robot.type=so100_follower \ --dataset.repo_id=<hf_user>/eval_<dataset> \ --policy.path=<hf_user>/<desired_policy_repo_id> \ --episodes=10 ``` Prefix the dataset repo with **eval\_** and supply `--policy.path` pointing to a local or hub checkpoint. --- ## Model Details - **License:** apache-2.0

Source context: 4 downloads · 0 likes · Pipeline robotics · Library lerobot · Repo chasedreaminf/pi05_so101-pick-block-in-cup-2cam-run1-v2

Source context: 0 downloads · 0 likes · Pipeline robotics · Library lerobot · Repo chasedreaminf/pi05_so101-pick-block-in-cup-2cam-run1-v2