distilbert-base-uncased-banking77-classification

<!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. -->

distilbert-base-uncased-banking77-classification

This model is a fine-tuned version of distilbert-base-uncased on the banking77 dataset. It achieves the following results on the evaluation set:

• Loss: 0.3152
• Accuracy: 0.9240
• F1 Score: 0.9243

Model description

This is my first fine-tuning experiment using Hugging Face. Using distilBERT as a pretrained model, I trained a classifier for online banking queries. It could be useful for addressing tickets.

Intended uses & limitations

The model can be used on text classification. In particular is fine tuned on banking domain.

Training and evaluation data

The dataset used is banking77

The 77 labels are:

Training procedure

from transformers import pipeline

pipe = pipeline("text-classification", model="nickprock/distilbert-base-uncased-banking77-classification")
pipe("I can't pay by my credit card")

Training hyperparameters

The following hyperparameters were used during training:

• learning_rate: 2e-05
• train_batch_size: 64
• eval_batch_size: 64
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 20

Training results

Framework versions

• Transformers 4.20.1
• Pytorch 1.12.0+cu113
• Datasets 2.3.2
• Tokenizers 0.12.1