Downloads · 30 days
0
realworld-q-recap-ps-sweep
sm-park/realworld-q-recap-ps-sweep
realworld-q-recap-ps-sweep is a robotics model from sm-park. 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 rlinf. The card lists the license as apache-2.0.
OpenPI π0.5 policies trained with Q-RECAP (CFG + 4-bit LM QAT) on real-world multitask data. Each subfolder is a different failure penalty with primitivesuccess shaping on the last step (failurereward - 500 (1 - p)).
Access
Public
Updated May 19, 2026
Repo size
51.2 GB
Likes
0
Public
Repo makeup
Click a slice to open those files.
.pt51.2 GB · 100%
At a glance
- Task
- Robotics
- Library
- rlinf
- License
- apache-2.0
- Access
- Public
- Created
- May 19, 2026
- Updated
- May 19, 2026
- SHA
- c0b3a128
Model card
From the Hugging Face model README
Real-world Q-RECAP policy sweep (fail*_ps)
OpenPI π0.5 policies trained with Q-RECAP (CFG + 4-bit LM QAT) on real-world multitask data.
Each subfolder is a different failure penalty with primitive_success shaping on the last step
(failure_reward - 500 * (1 - p)).
This repo contains policy checkpoints only (stage4 actor). Value / advantage artifacts are not included.
Layout
fail300_ps/actor/model_state_dict/full_weights.pt
fail500_ps/actor/model_state_dict/full_weights.pt
...
norm_stats.json # shared; also copied under each fail*_ps/
manifest.csv
Included runs
| Folder | failure_reward | training step |
|---|---|---|
fail300_ps | -300 | global_step_30000 |
fail400_ps | -400 | global_step_30000 |
fail500_ps | -500 | global_step_30000 |
fail600_ps | -600 | global_step_30000 |
fail750_ps | -750 | global_step_30000 |
fail1000_ps | -1000 | global_step_30000 |
Load in RLinf
export Q_RECAP_POLICY_PATH="/path/to/this-repo/fail500_ps/actor"
# eval / deploy with:
# examples/recap/cfg/config/realworld_cfg_openpi_q_recap_qat_lm4bit.yaml
# data.advantage_tag=realworld_q_recap_fail500_ps_N8_ckpt18000_q30
RLinf resolves actor/model_state_dict/full_weights.pt under model_path.
Base model
Fine-tuned from the real-world multitask π0.5 checkpoint used in this project
(see Q_RECAP_RAW_POLICY_PATH / Q_RECAP_POLICY_PATH in env.sh).
Confirm license terms for the base OpenPI / aiha weights before redistribution.
- Task
- Robotics
- Library
- rlinf
- License
- apache-2.0
- Created
- May 19, 2026
- Updated
- May 19, 2026