Skip to main content
Supervised Fine Tuning (SFT) lets you fine-tune and adapt Newton using your custom dataset. This is the most powerful customization option, producing a model specifically adapted to your sensor configuration or use case. The result of fine-tuning Newton is a forked instance of Newton that can be identified with a specific model_id. The fine-tuned version of Newton is encrypted and access-controlled with visibility to your organization only.

​
When to Use Fine-Tuning

Fine-tuning is appropriate when:
  • You have a unique use case or bespoke sensor that can benefit from custom training on a proprietary dataset
  • You have a substantial labeled dataset (typically hundreds or thousands of examples)
  • N-shot examples aren’t sufficient for your accuracy requirements
  • You need Newton to learn complex patterns specific to your domain

​
What Fine-Tuning Produces

The result of fine-tuning is a new Newton instance with:
  • A unique model_id that you use to access your custom model
  • Encryption and access controls limiting visibility to your organization only
  • Behavior adapted to your specific sensor configuration and use case based on the knowledge gained from fine tuning on your custom dataset.

​
Labeled Examples

SFT relies on high-quality examples to help steer a pre-trained instance of Newton to a new sensor configuration or use case. Each labeled example consists of an input and output pair in JSON or YAML format. This provides Newton the information required to understand that given input X, it should generate output Y. As Newton is a multimodal model, the input (X) can consist of one or more sensor types, with the output (Y) currently constrained to text output.

​
Example Formats

​
Sensor Log → Text

An example that converts sensor logs from motion sensors to a customer activity report: 
lens_parameters:
  instruction: |
    You are a retail monitoring system. Monitor the motion of customers 
    throughout a store based on the presence sensors.
    Output a report on the customers activity based on the logs in the 
    format: [time start - time end] {activity}.
inputs:
  - type: data.text
    event_data:
      contents: |
        - sensor_name: Entrance
          log:
            - presence_state: "active"
              timestamp: "2025-08-18T12:25:00"
            - presence_state: "inactive"
              timestamp: "2025-08-18T12:25:01"
        - sensor_name: Aisle1
          log:
            - presence_state: "inactive"
              timestamp: "2025-08-18T12:25:00"
        ...
outputs:
  - type: data.text
    event_data:
      contents: |
        Here is a summarized report of customer activity based on the given sensor logs.
        * [12:25:00 - 12:25:01] Customer at Entrance
        * [12:25:01 - 12:25:03] Customer in Aisle1
        * [12:25:03 - 12:25:06] Customer in Aisle2
        * [12:25:06 - 12:25:09] Customer in Aisle3
        * [12:25:09 - 12:25:10] Customer at Checkout
        * [12:25:10 - onward] Customer at Exit

​
Base64 Image → Text

An example that generates traffic alerts from camera images: 
lens_parameters:
  instruction: |
    You are an intelligent traffic monitoring system. Monitor the traffic 
    feed for the following activity or object.
    Output alerts in the format: [alert]{activity or object name}: {description}
    Context: this camera is on 104th and main street looking west-to-east 
    from left-to-right.
  focus: All red pickup trucks driving east.
inputs:
  - type: data.base64_img
    event_data:
      contents: iVBORw0KGgoAAAANSUh...
outputs:
  - type: data.text
    event_data:
      contents: "[alert]red pickup truck: A red pickup truck is driving on 104th and main heading east."

​
Data Event Types

The input and output data use Newton’s event protocol:
Event TypeDescription
data.textText content (sensor logs, reports, alerts)
data.base64_imgBase64-encoded image data
See Data Events for complete documentation on event types.

​
Best Practices

Quality over quantity A smaller set of high-quality, representative examples is more valuable than a large set of noisy or inconsistent examples. Consistent formatting Ensure your output format is consistent across examples. Newton will learn the patterns in your labeled data. Cover edge cases Include examples that represent the full range of scenarios Newton will encounter, including edge cases and unusual inputs.

​
Getting Started

To begin fine-tuning Newton, prepare your labeled dataset following the format shown above, then contact your Archetype representative to initiate the fine-tuning process.