Model Card for Model ID

This model is used to classify the user-intent for the Danswer project, visit https://github.com/danswer-ai/danswer.

Model Details

Multiclass classifier on top of distilbert-base-uncased

Model Description

Classifies user intent of queries into categories including: 0: Keyword Search 1: Semantic Search 2: Direct Question Answering

Developed by: [DanswerAI]
License: [MIT]
Finetuned from model [optional]: [distilbert-base-uncased]

Model Sources [optional]

Repository: [https://github.com/danswer-ai/danswer]
Demo [optional]: [Upcoming!]

Uses

This model is intended to be used in the Danswer Question-Answering System

Bias, Risks, and Limitations

This model has a very small dataset maintained by DanswerAI. If interested, reach out to danswer.dev@gmail.com.

Recommendations

This model is intended to be used in the Danswer (QA System)

How to Get Started with the Model

from transformers import AutoTokenizer
from transformers import TFDistilBertForSequenceClassification
import tensorflow as tf

model = TFDistilBertForSequenceClassification.from_pretrained("danswer/intent-model")
tokenizer = AutoTokenizer.from_pretrained("danswer/intent-model")

class_semantic_mapping = {
        0: "Keyword Search",
        1: "Semantic Search",
        2: "Question Answer"
    }

# Get user input
user_query = "How do I set up Danswer to run on my local environment?"

# Encode the user input
inputs = tokenizer(user_query, return_tensors="tf", truncation=True, padding=True)

# Get model predictions
predictions = model(inputs)[0]

# Get predicted class
predicted_class = tf.math.argmax(predictions, axis=-1)

print(f"Predicted class: {class_semantic_mapping[int(predicted_class)]}")