Sagemaker Roberta Base Emotion

🚀 sagemaker-roberta-base-emotion

This is a fine - tuned model for multi - class text classification (emotion detection), trained using Amazon SageMaker and the new Hugging Face Deep Learning container. It offers high - accuracy emotion detection capabilities.

🚀 Quick Start

This model is ready for use in text classification tasks, especially emotion detection. You can utilize it with the Hugging Face ecosystem and Amazon SageMaker.

✨ Features

• Fine - tuned: Based on the roberta - base model, fine - tuned for emotion detection.
• High Performance: Achieves high accuracy and F1 scores on the emotion dataset.

📦 Installation

No specific installation steps are provided in the original document.

💻 Usage Examples

No code examples are provided in the original document.

📚 Documentation

Model Details

• Model Name: sagemaker - roberta - base - emotion
• Problem Type: Multi Class Text Classification (emotion detection)

Results on Evaluation Set

Hyperparameters

{
    "epochs": 10,
    "train_batch_size": 16,
    "learning_rate": 3e-5, 
    "weight_decay":0.01,
    "load_best_model_at_end": true,
    "model_name":"roberta-base",
    "do_eval": True,
    "load_best_model_at_end":True
}

Validation Metrics

Model Index

• Name: sagemaker - roberta - base - emotion
• Results:
  • Task 1:
    • Task Type: text - classification
    • Task Name: Multi Class Text Classification
    • Dataset:
      • Name: emotion
      • Type: emotion
    • Metrics:
      • Accuracy: 94.1 (Validation Accuracy)
      • F1: 94.13 (Validation F1)
  • Task 2:
    • Task Type: text - classification
    • Task Name: Text Classification
    • Dataset:
      • Name: emotion
      • Type: emotion
      • Config: default
      • Split: test
    • Metrics:
      • Accuracy: 0.931 (verified)
      • Precision Macro: 0.8833042147663716 (verified)
      • Precision Micro: 0.931 (verified)
      • Precision Weighted: 0.9337002742192515 (verified)
      • Recall Macro: 0.9087144572668905 (verified)
      • Recall Micro: 0.931 (verified)
      • Recall Weighted: 0.931 (verified)
      • F1 Macro: 0.8949974527433656 (verified)
      • F1 Micro: 0.931 (verified)
      • F1 Weighted: 0.9318434300647934 (verified)
      • Loss: 0.17379647493362427 (verified)

🔧 Technical Details

The model is fine - tuned using Amazon SageMaker and the Hugging Face Deep Learning container. Hyperparameters are carefully selected to achieve good performance on the emotion dataset.

📄 License

This model is licensed under the Apache 2.0 license.