User Intention Recognition

🚀 User Intention Recognition

This project is designed for user intention recognition, leveraging a pre - trained model to classify various user intents in text conversations.

🚀 Quick Start

✨ Features

• Base Model: Utilizes google-bert/bert-base-uncased as the base model.
• Functionality: Specialized in text2text - generation for user intention recognition.

📦 Installation

The installation steps mainly involve installing the transformers library and torch. You can use the following command to install the transformers library:

pip install transformers

And install torch according to your CUDA environment:

pip install torch torchvision torchaudio

💻 Usage Examples

Basic Usage

from transformers import AutoTokenizer, AutoModelForSequenceClassification
import torch

model = AutoModelForSequenceClassification.from_pretrained("Savtale/User-Intention-Recognition")
tokenizer = AutoTokenizer.from_pretrained("Savtale/User-Intention-Recognition")

# User is talking with a chatbot
input_user_text = "Sounds good. I'm interested in trying the free trial. How do I sign up?"

# Tokenizing text
inputs = tokenizer(input_user_text.lower(), return_tensors="pt")

# Predict
with torch.no_grad():
    outputs = model(**inputs)
    logits = outputs.logits

# Prepare result
probabilities = torch.softmax(logits, dim=-1) # Top Match
predicted_class = int(torch.argmax(probabilities)) # Class no.
str_predicted_class = class_dict[str(predicted_class)] # Class String

print(f"Predicted User Intention: {str_predicted_class}") # Predicted User Intention: Request a Demo

Classes

class_dict = {
    "0": "Ask For Technical Support",
    "1": "Ask General Question",
    "2": "Start Conversation",
    "3": "Express Dissatisfaction",
    "4": "Request Product Information",
    "5": "Inquire About Pricing",
    "6": "Negotiate Price",
    "7": "Request Return or Refund",
    "8": "Provide Positive Feedback",
    "9": "Provide Negative Feedback",
    "10": "Seek Recommendation",
    "11": "Request Customization",
    "12": "Ask About Shipping and Delivery",
    "13": "Inquire About Warranty and Support",
    "14": "Express Interest in Upselling",
    "15": "Express Interest in Cross-selling",
    "16": "Request Urgent Assistance",
    "17": "Ask About Promotions and Discounts",
    "18": "Inquire About Loyalty Programs",
    "19": "Request a Callback",
    "20": "Ask About Payment Options",
    "21": "Express Uncertainty",
    "22": "Request Clarification",
    "23": "Confirm Understanding",
    "24": "End Conversation",
    "25": "Express Gratitude",
    "26": "Apologize",
    "27": "Complain About Customer Service",
    "28": "Request a Manager",
    "29": "Ask About Company Policies",
    "30": "Inquire About Job Opportunities",
    "31": "Ask About Corporate Social Responsibility",
    "32": "Express Interest in Investing",
    "33": "Cancellation",
    "34": "Ask About Return Policy",
    "35": "Inquire About Sustainability Practices",
    "36": "Request a Catalog",
    "37": "Ask About Brand History",
    "38": "Express Interest in Partnership",
    "39": "Inquire About Franchise Opportunities",
    "40": "Ask About Corporate Events",
    "41": "Express Interest in Volunteering",
    "42": "Request a Referral",
    "43": "Ask About Gift Cards",
    "44": "Inquire About Product Availability",
    "45": "Request a Personalized Recommendation",
    "46": "Ask About Order Status",
    "47": "Express Interest in a Webinar or Workshop",
    "48": "Request a Demo",
    "49": "Ask About Social Media Channels"
                }

📄 License

No license information is provided in the original document.

👥 Authors

• Savta