HP-BERT: A BERT-based model for quantifying subjective heat perception from Chinese social media texts. Includes the fine-tuned model, scoring script, and example data.
This repository provides the HP-BERT model and example code for quantifying subjective heat perception from Chinese social media texts. The ZIP file contains three main components: HP-BERT/ Fine-tuned HP-BERT model package. This folder contains the trained weights and configuration files required to load the HP-BERT model for heat perception prediction.
HP-BERT_score.py/ Inference script for scoring heat perception. This script loads the HP-BERT model and outputs heat perception scores for input texts. It can be used to reproduce the results in our paper or to apply HP-BERT to new social media data.
posts for testing.XLSX A sample set of 100 Weibo posts (anonymized) with varying levels of heat perception. This file is provided as a minimal example to show how to format input texts and how HP-BERT works in practice. 👉 [Download HP-BERT Model]:(https://github.com/PinggeHe/HP-BERT/releases/tag/v1.0.0)