> ## Documentation Index
> Fetch the complete documentation index at: https://docs.orchestrallabs.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# MolmoAct 2

> Inspect and deploy MolmoAct 2, with its camera and state inputs and action output.

## Checkpoint

```bash title="Your computer" sample=verified theme={null}
orchestral models inspect molmoact2 hf://lerobot/MolmoAct2-BimanualYAM-LeRobot@743af86d1b0ce3c95faa05b94573f9d067bd2e65
```

```text title="Sample output" theme={null}
Inputs:
  observation.images.top: RGB uint8 (224, 224, 3)
  observation.images.left: RGB uint8 (224, 224, 3)
  observation.images.right: RGB uint8 (224, 224, 3)
  state: float32 (14,)

Outputs:
  actions: float32 (30,14)
```

## Inputs and output

Use the camera keys listed below. Missing or extra keys raise an error.

The table describes the published BimanualYAM policy. For another checkpoint, use the camera keys and dimensions returned by `orchestral models inspect`.

| Direction | Name | Shape | Type |
| - | - | - | - |
| Input | `observation.images.top` | `(224, 224, 3)` | RGB `uint8` |
| Input | `observation.images.left` | `(224, 224, 3)` | RGB `uint8` |
| Input | `observation.images.right` | `(224, 224, 3)` | RGB `uint8` |
| Input | `state` | `(14,)` | `float32` |
| Output | `actions` | `(30, 14)` | `float32` |

## Deploy and predict

Deploy the checkpoint URI directly. Orchestral resolves, imports and prepares it before bringing up the serving deployment.

```bash your computer theme={null}
orchestral deployments create hf://lerobot/MolmoAct2-BimanualYAM-LeRobot@743af86d1b0ce3c95faa05b94573f9d067bd2e65 --robot robot-1 --wait
```

```text title="Sample output" theme={null}
jolly-tangerine-snail-rdik  ACTIVE · ready
MolmoAct2 · robot-1
Timeline: requested 12s ago → starting worker 8s ago → ready now
```

```python robot computer theme={null}
import numpy as np
import orchestral as ol

client = ol.Orchestral()
deployment_id = input("Deployment ID from the CLI output: ").strip()
deployment = client.deployments.get(deployment_id)

inputs = {
    "observation.images.top": np.zeros((224, 224, 3), dtype=np.uint8),
    "observation.images.left": np.zeros((224, 224, 3), dtype=np.uint8),
    "observation.images.right": np.zeros((224, 224, 3), dtype=np.uint8),
    "state": np.zeros((14,), dtype=np.float32),
}

with deployment.session() as session:
    result = session.predict(inputs=inputs, instruction="pick up the object")

print(result.actions.shape)  # (30, 14)
```

## Imported policies

Deploy the LeRobot policy checkpoint with its saved preprocessor and postprocessor files.
Inspection reads its declared camera features, state/action dimensions, and referenced backbone
checkpoint. Deployment pins the backbone revision and retains the policy weights and processors.


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.