Chinese Emotion

🚀 Chinese Text Emotion Classifier

This Chinese text emotion classifier is fine-tuned from a base model, featuring a smaller parameter size. It's an ideal choice for users who need faster inference speed, and its actual performance in tests is comparable to other models. Model: Chinese-Emotion-Small

🚀 Quick Start

Install Dependencies

Make sure you have installed Hugging Face's Transformers library and PyTorch:

pip install transformers torch

Load the Model

Use the following code to load the model and tokenizer, and perform emotion classification:

from transformers import AutoTokenizer, AutoModelForSequenceClassification
import torch

# Add device configuration
device = torch.device("cuda" if torch.cuda.is_available() else "cpu")

# Label mapping dictionary
label_mapping = {
    0: "Neutral tone",
    1: "Concerned tone",
    2: "Happy tone",
    3: "Angry tone",
    4: "Sad tone",
    5: "Questioning tone",
    6: "Surprised tone",
    7: "Disgusted tone"
}

def predict_emotion(text, model_path="Johnson8187/Chinese-Emotion"):
    # Load the model and tokenizer
    tokenizer = AutoTokenizer.from_pretrained(model_path)
    model = AutoModelForSequenceClassification.from_pretrained(model_path).to(device)  # Move the model to the device
    
    # Convert the text to the model input format
    inputs = tokenizer(text, return_tensors="pt", truncation=True, padding=True).to(device)  # Move the input to the device
    
    # Make predictions
    with torch.no_grad():
        outputs = model(**inputs)
    
    # Get the prediction result
    predicted_class = torch.argmax(outputs.logits).item()
    predicted_emotion = label_mapping[predicted_class]
    
    return predicted_emotion

if __name__ == "__main__":
    # Usage examples
    test_texts = [
        "Although I've tried hard for a long time, I always seem unable to do it. I feel completely useless.",
        "What you said really confused me. I have no idea how to react.",
        "The world is so heartless. Why does it always give me such tests?",
        "Sometimes, I just hope to have a little peace and stop hearing these boring topics.",
        "Every time I think of that past, my heart still aches. I really can't let it go.",
        "I never thought there would be such a big change. Now I feel completely out of control.",
        "I never expected you to do this. It surprised me beyond words.",
        "I know I should be stronger, but sometimes, these emotions really make me on the verge of collapse."
    ]

    for text in test_texts:
        emotion = predict_emotion(text)
        print(f"Text: {text}")
        print(f"Predicted emotion: {emotion}\n")

✨ Features

• Specializes in Chinese text emotion analysis.
• Can identify 8 emotion labels: Neutral tone, Concerned tone, Happy tone, Angry tone, Sad tone, Questioning tone, Surprised tone, Disgusted tone.
• Applicable to various scenarios such as customer service emotion monitoring, social media analysis, and user feedback classification.

📚 Documentation

Model Introduction

This model is fine-tuned based on the joeddav/xlm-roberta-large-xnli model, focusing on Chinese text emotion analysis. Through fine-tuning, the model can identify the following 8 emotion labels:

• Neutral tone
• Concerned tone
• Happy tone
• Angry tone
• Sad tone
• Questioning tone
• Surprised tone
• Disgusted tone

The model is suitable for various scenarios, including customer service emotion monitoring, social media analysis, and user feedback classification.

Dataset

• The fine-tuning dataset consists of 4,000 annotated Traditional Chinese emotion samples, covering various emotion categories to ensure the model's generalization ability in emotion classification.
• Johnson8187/Chinese_Multi-Emotion_Dialogue_Dataset

📄 License

This project is licensed under the MIT license.

🌟 Contact and Feedback

If you encounter any issues while using this model, please contact:

• Email: fable8043@gmail.com
• Hugging Face Project Page: chinese-text-emotion-classifier