Skip to content

Models calling functions

Model-Calling Functions

Model calling functions interface with external conversational or classification models. They prepare the input, send it to the model, and handle the response.

call_conv_model(instruction)

autograms.functional.call_conv_model Calls the conversational model with the specified instruction.

  • Parameters:

  • instruction (str): Instruction for the conversational model.

  • Returns:

  • str: The model's response.

call_classifier(input_str, answer_choices, model_type=None, model=None, **kwargs)

autograms.functional.call_classifier Calls a classification model to determine the best choice.

  • Parameters:
  • input_str (str): The input text for classification.
  • answer_choices (list[str]): Possible answers.
  • model_type (str, optional): Type of the model.
  • model (str, optional): Path to the model.
  • multi_modal_inputs (list[dict], optional) --list of open ai style image or audio inputs for the model, for example : 
    {
"type": "image_url",
"image_url": {
    "url": "https://upload.wikimedia.org/wikipedia/commons/thumb/d/dd/Gfp-wisconsin-madison-the-nature-boardwalk.jpg/2560px-Gfp-wisconsin-madison-the-nature-boardwalk.jpg"}
}

  • **kwargs: Additional model arguments.

  • Returns:

  • tuple: The chosen answer and a success flag.

call_model(input_turns, output_turns, system_prompt, system_prompt_in_turns=False, model_type=None, model=None, **kwargs)

autograms.functional.call_model Calls a conversational model and retrieves a response.

  • Parameters:
  • input_turns (list): The conversation input turns.
  • output_turns (list): The conversation output turns.
  • system_prompt (str): System prompt to guide the model.
  • system_prompt_in_turns (bool, default=False): Whether the system prompt is included in the turns.
  • model_type (str, optional): The type of model to call.
  • model (str, optional): Path to a specific model.
  • multi_modal_inputs (list[dict], optional) --list of open ai style image or audio inputs for the model, for example : 
    {
"type": "image_url",
"image_url": {
    "url": "https://upload.wikimedia.org/wikipedia/commons/thumb/d/dd/Gfp-wisconsin-madison-the-nature-boardwalk.jpg/2560px-Gfp-wisconsin-madison-the-nature-boardwalk.jpg"}
}

  • **kwargs: Additional arguments for the model.

  • Returns:

  • tuple: A tuple containing:
    • result (str): The model response.
    • success (bool): Whether the model call was successful.

call_object_model(input_turns, output_turns, system_prompt, system_prompt_in_turns=False, model_type=None, model=None, obj_structure=None, **kwargs)

autograms.functional.call_object_model Calls a model to generate a structured object based on input.

  • Parameters:
  • input_turns (list[dict]): User input history.
  • output_turns (list[dict]): Model output history.
  • system_prompt (str): The system prompt.
  • system_prompt_in_turns (bool): Include system prompt in turns.
  • model_type (str, optional): Model type.
  • model (str, optional): Model path.
  • obj_structure (BaseModel): Pydantic model structure or json schema.
  • multi_modal_inputs (list[dict], optional) --list of open ai style image or audio inputs for the model, for example : 
    {
"type": "image_url",
"image_url": {
    "url": "https://upload.wikimedia.org/wikipedia/commons/thumb/d/dd/Gfp-wisconsin-madison-the-nature-boardwalk.jpg/2560px-Gfp-wisconsin-madison-the-nature-boardwalk.jpg"}
}

  • **kwargs: Additional arguments.

  • Returns:

  • BaseModel: Generated object.