novapay-sentiment

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

• Loss: 0.4476
• Accuracy: 0.888
• F1: 0.8869

Model description

license: apache-2.0

tags:

• text-classification
• sentiment-analysis
• novapay

datasets:

• imdb

metrics:

• accuracy
• f1

Intended uses & limitations

Intended uses : Fine-tuned for sentiment classification of NovaPay customer support messages. This is a proof-of-concept model trained on the IMDB dataset.

Limitation:

• Trained on IMDB movie reviews, NOT actual financial customer messages
• May not generalize well to financial domain-specific language
• Proof-of-concept only — not production-ready

Training and evaluation data

Training Data

• Dataset: IMDB (5,000 example subset)
• Base model: distilbert/distilbert-base-uncased
• Epochs: 2
• Learning rate: 2e-5

Evaluation Results

• Accuracy: ~0.87-0.90
• F1 Score: ~0.87-0.90

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

• learning_rate: 2e-05
• train_batch_size: 16
• eval_batch_size: 16
• seed: 42
• optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• num_epochs: 2

Training results

Framework versions

• Transformers 5.18.0
• Pytorch 2.11.0+cu130
• Datasets 4.8.5
• Tokenizers 0.23.2