Bert Gomotions

🚀 🔥 在GoEmotions數據集上微調的BERT模型

本模型是在GoEmotions數據集上對BERT進行微調，用於多標籤情感分類。它能夠為輸入的文本預測多種情感。

🚀 快速開始

模型使用

from transformers import AutoTokenizer, AutoModelForSequenceClassification
import torch

# 加載模型和分詞器
model_name = "codewithdark/bert-Gomotions"
tokenizer = AutoTokenizer.from_pretrained(model_name)
model = AutoModelForSequenceClassification.from_pretrained(model_name)

# 情感標籤（根據你的數據集進行調整）
emotion_labels = [
    "Admiration", "Amusement", "Anger", "Annoyance", "Approval", "Caring", "Confusion",
    "Curiosity", "Desire", "Disappointment", "Disapproval", "Disgust", "Embarrassment",
    "Excitement", "Fear", "Gratitude", "Grief", "Joy", "Love", "Nervousness", "Optimism",
    "Pride", "Realization", "Relief", "Remorse", "Sadness", "Surprise", "Neutral"
]

# 示例文本
text = "I'm so happy today!"
inputs = tokenizer(text, return_tensors="pt")

# 預測
with torch.no_grad():
    outputs = model(**inputs)
    probs = torch.sigmoid(outputs.logits).squeeze(0)  # 將logits轉換為概率

# 獲取前5個預測結果
top5_indices = torch.argsort(probs, descending=True)[:5]  # 獲取前5個標籤的索引
top5_labels = [emotion_labels[i] for i in top5_indices]
top5_probs = [probs[i].item() for i in top5_indices]

# 打印結果
print("Top 5 Predicted Emotions:")
for label, prob in zip(top5_labels, top5_probs):
    print(f"{label}: {prob:.4f}")

'''
輸出:
Top 5 Predicted Emotions:
Joy: 0.9478
Love: 0.7854
Optimism: 0.6342
Admiration: 0.5678
Excitement: 0.5231
'''

在Hugging Face中使用

from transformers import pipeline

classifier = pipeline("text-classification", model="codewithdark/bert-Gomotions", top_k=None)
classifier("I'm so excited about the trip!")

✨ 主要特性

• 這是一個微調版的BERT（bert-base-uncased）模型，在GoEmotions數據集上進行訓練，可用於多標籤情感分類，能為每個輸入文本預測多種情感。

📊 性能

🔧 技術細節

• 模型：bert-base-uncased
• 數據集：GoEmotions
• 優化器：AdamW
• 損失函數：BCEWithLogitsLoss（用於多標籤分類的二元交叉熵）
• 批量大小：16
• 訓練輪數：3
• 評估指標：準確率、F1分數、漢明損失

📄 許可證

本模型使用的許可證為 apache-2.0。

🛠️ 引用

如果您使用了此模型，請引用以下內容：

@misc{your_model,
  author = {codewithdark},
  title = {Fine-tuned BERT on GoEmotions},
  year = {2025},
  url = {https://huggingface.co/codewithdark/bert-Gomotions}
}