from transformers import AutoTokenizer, AutoModelForSequenceClassification
import torch

# Load model and tokenizer
tokenizer = AutoTokenizer.from_pretrained('whitedevil0089devil/roberta_base_1')
model = AutoModelForSequenceClassification.from_pretrained('whitedevil0089devil/roberta_base_1')

# Example usage
question = "Your question here"
inputs = tokenizer(question, return_tensors="pt", truncation=True, padding=True, max_length=384)

with torch.no_grad():
    outputs = model(**inputs)
    predictions = torch.nn.functional.softmax(outputs.logits, dim=-1)
    predicted_class = torch.argmax(outputs.logits, dim=-1).item()
    confidence = predictions[0][predicted_class].item()

print(f"Predicted class: {predicted_class}")
print(f"Confidence: {confidence:.4f}")

Training Details

This model was fine-tuned using:

Framework: PyTorch + Transformers
Optimization: AdamW with learning rate scheduling
Training Strategy: Early stopping with validation monitoring
Hardware: Trained on Google Colab (T4 GPU)

Intended Use

This model is designed for question-answering classification tasks. It can be used to:

Classify questions into predefined categories
Provide automated responses based on question classification
Support Q&A systems and chatbots

Limitations

Model performance depends on the similarity between training data and inference data
May not generalize well to domains significantly different from training data
Classification accuracy may vary based on question complexity and length

Citation

If you use this model, please cite:

@misc{roberta-qa-model,
  title={Fine-tuned RoBERTa for Question-Answer Classification},
  author={Your Name},
  year={2024},
  url={https://huggingface.co/whitedevil0089devil/roberta_base_1}
}