> ## 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.

# Isaac 0.5

> Inspect and deploy Isaac 0.5, with its camera and state inputs and action output.

## Checkpoint

```bash title="Your computer" sample=verified theme={null}
orchestral models inspect isaac05 hf://PerceptronAI/Isaac-0.5@6e397429f0ab91528ebac915004eba5e763fcdd4
```

```text title="Sample output" theme={null}
Inputs:
  observation.images.image: RGB uint8 (256, 256, 3)
  observation.images.wrist_image: RGB uint8 (256, 256, 3)
  state: float32 (8,)

Outputs:
  actions: float32 (8,7)
```

## Inputs and output

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

The bundle uses the inputs from Isaac’s reference server because this checkpoint’s saved LeRobot configuration has empty feature dictionaries. A checkpoint with different declared features fails to load.

| Direction | Name | Shape | Type |
| - | - | - | - |
| Input | `observation.images.image` | `(256, 256, 3)` | RGB `uint8` |
| Input | `observation.images.wrist_image` | `(256, 256, 3)` | RGB `uint8` |
| Input | `state` | `(8,)` | `float32` |
| Output | `actions` | `(8, 7)` | `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://PerceptronAI/Isaac-0.5@6e397429f0ab91528ebac915004eba5e763fcdd4 --robot robot-1 --wait
```

```text title="Sample output" theme={null}
jolly-tangerine-snail-rdik  ACTIVE · ready
Isaac 0.5 · 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.image": np.zeros((256, 256, 3), dtype=np.uint8),
    "observation.images.wrist_image": np.zeros((256, 256, 3), dtype=np.uint8),
    "state": np.zeros((8,), dtype=np.float32),
}

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

print(result.actions.shape)  # (8, 7)
```

## Imported policies

Inspection reads the saved LeRobot policy configuration. Camera names, image size, state
width, and action dimensions come from that configuration. The server runs the saved upstream
preprocessor and postprocessor.


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