Kayvane's picture
Update README.md
bacfe26
|
raw
history blame
1.68 kB
metadata
tags:
  - generated_from_trainer
datasets:
  - consumer_complaints
model-index:
  - name: distilbert-complaints-product
    results: []

distilbert-complaints-product

This model was trained from the (CFBP)[https://www.consumerfinance.gov/data-research/consumer-complaints/] dataset, also made available on the HuggingFace Datasets library. This model predicts the type of financial complaint based on the text provided

Model description

A DistilBert Text Classification Model, with 18 possible classes to determine the nature of a financial customer complaint.

Intended uses & limitations

This model is used as part of.a demonstration for E2E Machine Learning Projects focused on Contact Centre Automation:

  • Infrastructure: Terraform
  • MLOps: HuggingFace (Datasets, Hub, Transformers)
  • Cloud: AWS
    • Model Hosting: Lambda
    • DB Backend: DynamoDB
    • Orchestration: Step-Functions
    • UI Hosting: EC2
    • Routing: Gateway
  • UI: Budibase

Training and evaluation data

consumer_complaints dataset

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

  • learning_rate: 5e-05
  • train_batch_size: 32
  • eval_batch_size: 64
  • seed: 42
  • optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • lr_scheduler_type: linear
  • lr_scheduler_warmup_steps: 500
  • num_epochs: 3

Framework versions

  • Transformers 4.16.1
  • Pytorch 1.10.0+cu111
  • Datasets 1.18.2
  • Tokenizers 0.11.0