---
license: apache-2.0
datasets:
- XiangPan/waimai_10k
language:
- zh
metrics:
- accuracy
base_model:
- google-bert/bert-base-chinese
---
# Introduction

This model is trained based on the **base_model:google-bert/bert-base-chinese** and **datasets:XiangPan/waimai_10k** for sentiment analysis of reviews on a food delivery platform. It is designed to quickly identify negative reviews, allowing merchants to make targeted improvements efficiently.

# How to use

```bash
from transformers import AutoModelForSequenceClassification, AutoTokenizer
import torch

# 设备设置
device = torch.device("cuda" if torch.cuda.is_available() else "cpu")

# 加载预训练的模型和分词器
model_name = "zzz16/Public-analysis"  # 确保该模型路径正确
tokenizer_name = "bert-base-chinese"
model = AutoModelForSequenceClassification.from_pretrained(model_name).to(device)
tokenizer = AutoTokenizer.from_pretrained(tokenizer_name)

# 输入文本
text = "这个外卖平台的服务很差劲，配送慢，食物也冷了。"

# 使用分词器进行编码，将文本转化为模型输入的格式
inputs = tokenizer(text, padding=True, truncation=True, return_tensors="pt")
inputs = {key: value.to(device) for key, value in inputs.items()}  # 迁移到设备上

# 使用模型进行预测
with torch.no_grad():
    outputs = model(**inputs)

# 获取模型的输出结果
logits = outputs.logits
predicted_class = torch.argmax(logits, dim=-1)

# 输出预测的类别
print(f"预测类别: {predicted_class.item()}")
```

# 合作

我们在研发针对商家/企业/平台的外卖、舆情分析部署，主要针对商家/企业/平台进行舆情把控、情感分析，以进行针对性、快速应对和解决问题，如果您的公司想要体验或者是合作可以联系我们：3022656072@qq.com **邮件最好用中文！英文垃圾邮件太多，可能会回复不及时**