📘 Overview

PairQualSampleEvaluator is a text quality scoring operator based on the BGE model, trained on GPT pairwise comparison data, supporting both English and Chinese.

__init__ function

def __init__(self, model_cache_dir:str='./dataflow_cache', device="cuda", lang='en', max_length=512)

Parameter	Type	Default Value	Description
model_cache_dir	str	'./dataflow_cache'	The directory to cache the downloaded model.
device	str	"cuda"	The device to run the model on (e.g., "cuda" or "cpu").
lang	str	'en'	The language of the model to use. Supports 'en' or 'zh'.
max_length	int	512	The maximum sequence length for the tokenizer.

Prompt Template Descriptions

Prompt Template Name	Primary Use	Applicable Scenarios	Feature Description

run function

def run(self, storage: DataFlowStorage, input_key: str, output_key: str='PairQualScore')

Parameter	Type	Default Value	Description
storage	DataFlowStorage	Required	The data flow storage instance for reading and writing data.
input_key	str	Required	The name of the input column containing the text to be evaluated.
output_key	str	'PairQualScore'	The name of the output column where the quality scores will be stored.

🧠 Example Usage

🧾 Default Output Format

Field	Type	Description
input_key	str	The original input text from the specified input column.
PairQualScore	float	The calculated quality score, ranging from 0 to 1.

Example Input:

{
"instruction":"This is a high-quality piece of text that should receive a good score."
}

Example Output:

{
"instruction":"This is a high-quality piece of text that should receive a good score.",
"PairQualScore": 0.958
}