BERT-BankingClassifier

This model is a fine-tuned version of google-bert/bert-base-uncased on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.2867
• Accuracy: 0.9326

Model description

This model is a BERT (bert-base-uncased) model fine-tuned using the Banking77 dataset for text classification tasks. Banking77 is a dataset designed to categorize customer support messages in the banking and finance domain into 77 distinct categories. This model is specifically tailored for classifying customer complaints and inquiries in the banking sector, helping to assign messages to the correct categories effectively. Fine-Tuning Process : https://github.com/saribasmetehan/bank_administrative_assistant/blob/main/BERT_BankingClassifier.ipynb

Example

from transformers import pipeline, AutoTokenizer

text = "i forgot my PIN"
model_id = "saribasmetehan/BERT-BankingClassifier"

classifier = pipeline("text-classification", model=model_id)

preds = classifier(text)
print(preds)

#[{'label': 'LABEL_27', 'score': 0.9552194476127625}]

Training hyperparameters

The following hyperparameters were used during training:

• learning_rate: 5e-05
• train_batch_size: 16
• eval_batch_size: 16
• seed: 42
• optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• num_epochs: 4

Training results

Framework versions

• Transformers 4.46.3
• Pytorch 2.5.1+cu121
• Datasets 3.1.0
• Tokenizers 0.20.3