📘 Overview

The Phi4QAGenerator is an operator designed to generate multi-turn conversational question-and-answer data in a pre-training format based on given document content. It transforms raw text into a dialogue format suitable for language model pre-training.

__init__

def __init__(self, llm_serving: LLMServingABC)

Parameter	Type	Default	Description
llm_serving	LLMServingABC	Required	An LLM serving object that implements the LLMServingABC interface for model inference.

Prompt Template Descriptions

Prompt Template Name	Primary Use	Applicable Scenarios	Feature Description

run

def run(self, storage: DataFlowStorage, input_key: str = "raw_content", output_key: str = "generated_content")

Parameter	Type	Default	Description
storage	DataFlowStorage	Required	The data flow storage instance used for reading the input DataFrame and writing the output.
input_key	str	"raw_content"	The column name in the input DataFrame that contains the source document content.
output_key	str	"generated_content"	The column name where the generated conversational data will be stored.

🧠 Example Usage

# Blank

🧾 Output Format

Field	Type	Description
raw_content	str	The original input document text.
generated_content	str	The generated multi-turn dialogue suitable for pre-training.

Example Input:

{
  "raw_content": "The Eiffel Tower is a wrought-iron lattice tower on the Champ de Mars in Paris, France. It is named after the engineer Gustave Eiffel, whose company designed and built the tower."
}

Example Output:

{
  "raw_content": "The Eiffel Tower is a wrought-iron lattice tower on the Champ de Mars in Paris, France. It is named after the engineer Gustave Eiffel, whose company designed and built the tower.",
  "generated_content": "<|user|>\nWhat is the Eiffel Tower made of?\n<|end|>\n<|assistant|>\nThe Eiffel Tower is a wrought-iron lattice tower.\n<|end|>\n<|user|>\nWho is it named after?\n<|end|>\n<|assistant|>\nIt is named after the engineer Gustave Eiffel, whose company designed and built the tower.\n<|end|>"
}