liudev/chinese-roberta-novel-regression

Chinese-RoBERTa-WWM-Ext Novel Regressor

基于 hfl/chinese-roberta-wwm-ext 微调的文本回归模型，用于小说内容的情绪量化。

模型详情

• Base Model: hfl/chinese-roberta-wwm-ext
• 任务: 回归分析 (Regression)
• 输入长度: 512 tokens

如何使用

from transformers import AutoModelForSequenceClassification, AutoTokenizer
model_name = "liudev/chinese-roberta-novel-regression"
tokenizer = AutoTokenizer.from_pretrained(model_name)
model = AutoModelForSequenceClassification.from_pretrained(model_name)
text = "小明很生气，后果很严重"
inputs = tokenizer(text, return_tensors="pt", truncation=True, max_length=512)
outputs = model(**inputs)
# 获取回归分数
score = outputs.logits.item()
print(f"预测得分: {score}")
# 预测得分: -0.7371920347213745