Model Card for Sentiment Analysis (Positive / Negative / Neutral)
Model Details
Model Description
This model is a fine-tuned BERT-based transformer model for sentiment analysis.
It is trained using the GoEmotions dataset, where the original 27 emotion labels are mapped into three categories:
- Positive 😊
- Negative 😡
- Neutral 😐
The model takes text input and predicts its overall sentiment.
- Developed by: Krish Agrawal
- Model type: BERT (Transformer-based classification model)
- Language(s): English
- License: Apache 2.0
- Finetuned from model: bert-base-uncased
Model Sources
- Repository: https://github.com/krishagrawal623/bert-goemotions-sentiment-model.git
- Dataset: GoEmotions (Google Research)
Uses
Direct Use
This model can be used for:
- Sentiment analysis (positive / negative / neutral)
- Social media monitoring
- Customer feedback analysis
- Review classification
Example:
- Input: "This product is amazing!"
- Output: Positive
Downstream Use
- Chatbots 🤖
- Business analytics dashboards
- Customer support systems
- Market research tools
Out-of-Scope Use
- Not suitable for:
- Non-English text
- Detecting detailed emotions (like anger, joy, fear separately)
- Sarcasm or complex context
Bias, Risks, and Limitations
- Mapping 27 emotions → 3 classes may lose detailed emotional information
- May misclassify:
- Sarcasm
- Mixed sentiments
- Dataset bias may affect predictions
Recommendations
- Use only for general sentiment analysis
- Avoid using for sensitive or critical decisions
- Fine-tune further for domain-specific tasks
How to Get Started with the Model
from transformers import pipeline
classifier = pipeline("text-classification", model="your-username/your-model-name")
result = classifier("I am very happy today!")
print(result)
- Downloads last month
- 4