qhchina/SikuBERT-parallelism-wuyan-0.1

SikuBERT Parallel Couplet Classifier (五言对联平行性分类器)

A fine-tuned BERT model for detecting parallelism in classical Chinese couplets (五言对联), based on SIKU-BERT/sikubert.

Model Description

• Architecture: SIKU-BERT/sikubert fine-tuned for sequence classification
• Task: Binary classification (parallel vs non-parallel couplets)
• Labels:
  • parallel (对仗工整)
  • non-parallel (非对仗)

Usage

Direct Classification with Pipeline

from transformers import pipeline

# Load the classification pipeline
classifier = pipeline(
    "text-classification",
    model="qhchina/SikuBERT-parallelism-wuyan-0.1",
    tokenizer="qhchina/SikuBERT-parallelism-wuyan-0.1"
)

# Example parallel couplet
parallel_couplet = "感时花溅泪，恨别鸟惊心"

# Classify the example
result = classifier(parallel_couplet)

print("Classification result:")
print(f"Text: {parallel_couplet}")
print(f"Label: {result[0]['label']}")
print(f"Score: {result[0]['score']:.4f}")

# Text: 感时花溅泪，恨别鸟惊心
# Label: parallel
# Score: 0.9920

# Example non-parallel couplet
non_parallel_couplet = "看花随节序，不敢强为容"

# Classify the example
result = classifier(non_parallel_couplet)

print("Classification result:")
print(f"Text: {non_parallel_couplet}")
print(f"Label: {result[0]['label']}")
print(f"Score: {result[0]['score']:.4f}")

# Text: 看花随节序，不敢强为容
# Label: non-parallel
# Score: 0.9953