BERT-Banking77 Open-Source Bank Customer Service Text Classification Model

BERT Banking77

Developed by philschmid

A BERT-based banking customer service text classification model trained on the BANKING77 dataset with an accuracy of 92.64%

Text Classification

Transformers

English#Banking Customer Service Classification #High-Precision Text Classification #Financial Domain Specific

Downloads 74

Release Time : 6/2/2022

Model Overview

This model is specifically designed for text intent classification in banking customer service scenarios, accurately identifying 77 different types of banking-related user queries.

Model Features

High Accuracy

Achieves 92.64% accuracy and F1 score on the BANKING77 test set

Domain-Specific

Optimized specifically for banking customer service scenarios, covering 77 common banking service intents

Efficient Inference

Based on BERT architecture, enabling fast inference while maintaining high performance

Model Capabilities

Banking Customer Service Text Classification

Intent Recognition

Multi-Class Classification

Use Cases

Banking Customer Service

Credit Card Query Classification

Automatically classifies user queries about credit cards, such as application status, billing issues, etc.

Accurately identifies over 92% of credit card-related queries

Account Issue Routing

Automatically routes user account-related issues to the correct department

Reduces manual classification workload and improves customer service efficiency

🚀 BERT-Banking77 Model Trained Using AutoTrain

This model is designed for multi - class text classification on the Banking77 dataset, offering high - accuracy results with low CO2 emissions.

🚀 Quick Start

You can quickly start using the BERT - Banking77 model through the following methods:

Using cURL

$ curl -X POST -H "Authorization: Bearer YOUR_API_KEY" -H "Content-Type: application/json" -d '{"inputs": "I love AutoTrain"}' https://api-inference.huggingface.co/models/philschmid/autotrain-does-it-work-940131041

Using Python API

from transformers import AutoTokenizer, AutoModelForSequenceClassification, pipeline
model_id = 'philschmid/BERT-Banking77'
tokenizer = AutoTokenizer.from_pretrained(model_id)
model = AutoModelForSequenceClassification.from_pretrained(model_id)
classifier = pipeline('text-classification', tokenizer=tokenizer, model=model)
classifier('What is the base of the exchange rates?')

✨ Features

Multi - class Classification: Capable of handling multi - class text classification tasks, especially suitable for the Banking77 dataset.
Low CO2 Emissions: With a CO2 emission of only 0.03330651014155927 grams during training, it is environmentally friendly.

📚 Documentation

Model Information

Property	Details
Model Type	BERT - Banking77
Training Data	banking77
Problem Type	Multi - class Classification
Model ID	940131041
CO2 Emissions (in grams)	0.03330651014155927

Validation Metrics

Loss: 0.3505457043647766
Accuracy: 0.9263261296660118
Macro F1: 0.9268371013605569
Micro F1: 0.9263261296660118
Weighted F1: 0.9259954221865809
Macro Precision: 0.9305746406646502
Micro Precision: 0.9263261296660118
Weighted Precision: 0.929031563971418
Macro Recall: 0.9263724620088746
Micro Recall: 0.9263261296660118
Weighted Recall: 0.9263261296660118

Model Index Results

Task 1:
- Task Type: Text Classification
- Dataset: BANKING77
- Metrics:
  - Accuracy: 92.64
  - Macro F1: 92.64
  - Weighted F1: 92.6
Task 2:
- Task Type: Text Classification
- Dataset: banking77 (test split)
- Metrics:
  - Accuracy: 0.9275974025974026
  - Precision Macro: 0.9305185253845069
  - Precision Micro: 0.9275974025974026
  - Precision Weighted: 0.9305185253845071
  - Recall Macro: 0.9275974025974028
  - Recall Micro: 0.9275974025974026
  - Recall Weighted: 0.9275974025974026
  - F1 Macro: 0.927623314966026
  - F1 Micro: 0.9275974025974026
  - F1 Weighted: 0.927623314966026
  - Loss: 0.3199225962162018

Featured Recommended AI Models

Empowering the Future, Your AI Solution Knowledge Base

English 简体中文繁體中文にほんご