<< Click to Display Table of Contents >> Navigation: 3. Script Language > AI - Artificial Intelligence Commands > AIC. - Artificial Intelligence Command > Set_Model Commands > Set Model Commands |
MiniRobotLanguage (MRL)
AIC.Set Vision Detail
Configures the level of detail for image processing in OpenAI's vision API. It adjusts how the model processes the image and generates its textual understanding.
Intention
This command is intended to offer users control over the fidelity of image analysis – whether to prioritize speed and efficiency with a lower level of detail, or to seek a more detailed analysis at the expense of more resources and time.
•Low Fidelity (1 or "Low"):
•In this mode, the model receives a low-resolution 512 x 512 version of the image and represents it with a budget of 65 tokens.
•This mode is optimal for faster responses and fewer token consumption, suitable for use cases not requiring high detail.
•High Fidelity (2 or "High"):
•This mode enables detailed image processing. Initially, the model sees the low-resolution image, followed by detailed crops of the input images as 512px squares.
•Each detailed crop uses a token budget of 65 tokens, amounting to a total of 129 tokens.
•Opt for this mode when detailed image analysis is required.
Syntax
AIC.Set Vision Detail[|P1]
Parameter Explanation
P1 - (optional) is the level of detail setting for image processing. It can be set to:
•Low Fidelity (1 or "Low"):
•In this mode, the model receives a low-resolution 512 x 512 version of the image and represents it with a budget of 65 tokens.
•This mode is optimal for faster responses and fewer token consumption, suitable for use cases not requiring high detail.
•High Fidelity (2 or "High"):
•This mode enables detailed image processing. Initially, the model sees the low-resolution image, followed by detailed crops of the input images as 512px squares.
•Each detailed crop uses a token budget of 65 tokens, amounting to a total of 129 tokens.
•Opt for this mode when detailed image analysis is required.
Example
'***********************************
' AIC.-Sample
'***********************************
'Example SPR-Script to
' set vision detail to high fidelity
AIC.Set Vision Detail|2
' Or, to set it to low fidelity
AIC.Set Vision Detail|1
Remarks
Selecting the appropriate detail level is crucial based on the specific requirements of your image processing task. Higher detail provides more comprehensive analysis but consumes more resources.
The choice of detail level directly impacts the response time and token consumption of the AI model.
The Chat Completions API, being non-stateful, requires the image to be passed each time a request is made.
Limitations:
-
See also:
•