README.md

metadata

title: NLP Evaluation API
emoji: 🤖
colorFrom: blue
colorTo: green
sdk: docker
pinned: false

NLP Evaluation API

Multi-task NLP model for interview evaluation using RoBERTa-base (126M parameters).

Features

• Sentiment Analysis: negative, neutral, positive
• Emotion Detection: nervous, confident, stressed, calm, motivated
• Communication Quality: poor, fair, good, excellent
• Confidence Level: low, medium, high
• Stress Level: low, medium, high

API Endpoints

POST /analyze

Analyze text and get NLP predictions.

Request:

{
  "text": "I have extensive experience with React and have built scalable applications."
}

Response:

{
  "sentiment": "positive",
  "emotion": "confident",
  "communication": "excellent",
  "confidence_level": "high",
  "stress_level": "low",
  "scores": {
    "sentiment": 0.95,
    "emotion": 0.92,
    "communication": 0.94,
    "confidence_level": 0.91,
    "stress_level": 0.93
  }
}

Model Details

• Architecture: RoBERTa-base with multi-task learning
• Parameters: 126M
• Training: 8,000 samples, 3 epochs
• Accuracy: 92-96%