---
license: apache-2.0
tags:
  - robotics
  - vla
  - pi0.5
  - openpi
  - real-robot
  - franka
  - mode-editing
---

# pi05_real_pk_mixed

Fine-tuned [pi0.5](https://github.com/Physical-Intelligence/openpi) vision-language-action (VLA) model for real robot manipulation.

## Task

- **Task:** Pass Knife
- **Training data:** Mixed modes (left-hand + right-hand + sharp-end)
- **Dataset:** `real_pass_knife_mixed`
- **Robot:** Franka Panda (7-DOF)
- **Cameras:** Base RGB + Wrist RGB (256x256)

## Training Configuration

| Parameter | Value |
|-----------|-------|
| Base model | pi0.5 (PaliGemma 2B + Gemma 2B action expert) |
| Total parameters | ~3.35B |
| Action dimension | 32 |
| Action horizon | 10 |
| Batch size | 16 |
| Training steps | 5,000 |
| Learning rate | Cosine decay: warmup=500, peak=5e-5, end=5e-6 |
| Optimizer | AdamW (gradient clip norm=1.0) |
| Base weights | `gs://openpi-assets/checkpoints/pi05_base/params` |
| GPUs | 8x NVIDIA A100 |
| Normalization | Quantile normalization |

## Included Checkpoints

- **Step 4500**: loss = 0.0037
- **Step 4999**: loss = 0.0035

## Loss Curve

| Step | Loss |
|------|------|
| 0 | 0.0837 |
| 500 | 0.0155 |
| 1000 | 0.0133 |
| 1500 | 0.0107 |
| 2000 | 0.0084 |
| 2500 | 0.0078 |
| 3000 | 0.0068 |
| 3500 | 0.0055 |
| 4000 | 0.0043 |
| 4500 | 0.0037 |
| 4900 | 0.0035 |

## Usage with openpi

```python
# Add config name to openpi training config, then:
from openpi.training.config import get_config
config = get_config("pi05_real_pk_mixed")

# For inference, load the params checkpoint:
# checkpoint_path = "path/to/step_XXXX/params"
```

## Part of Mode Editing Research

This checkpoint is part of the **"Don't Filter Your Data, Edit Your Policy"** project (CoRL 2026), investigating post-hoc behavior mode editing for robot policies using Classifier-Guided Distillation (CG-Distill).

- **Mixed models** are trained on demonstrations containing all behavioral modes
- **Mode-specific models** are trained on single-mode filtered data
- **CG-Distill edited models** (coming soon) use classifier gradients to steer mixed models toward specific modes at zero inference cost