Fine-tuned DistilBERT-base-uncased for IMDB Classification

Model Description

DistilBERT is a transformer model that performs sentiment analysis. I fine-tuned the model on IMDB dataset with the purpose of classifying positive reviews from the bad ones. The model predicts these 2 classes.

The model is a fine-tuned version of DistilBERT.

It was fine-tuned on IMDB dataset [https://huggingface.co/datasets/imdb].

This model is a fine-tuned version of distilbert-base-uncased on IMDB dataset.

It achieves the following results on the evaluation set:

Loss: 0.2265
Accuracy: 0.9312

Intended uses & limitations

More information needed

Training and evaluation data

More information needed

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: Adam with betas=(0.9,0.999) and epsilon=1e-08
lr_scheduler_type: linear
num_epochs: 2

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
0.2273	1.0	1563	0.2471	0.9122
0.1524	2.0	3126	0.2265	0.9312

Framework versions

Transformers 4.35.2
Pytorch 2.1.0+cu121
Datasets 2.17.0
Tokenizers 0.15.1

How to Use

from transformers import pipeline

classifier = pipeline("text-classification", model="LukeGPT88/imdb_text_classifier")
classifier("I see it and it was awesome.")

Output:
[{'label': 'POSITIVE', 'score': 0.9958052635192871}]

Contact

Please reach out to luca.flammia@gmail.com if you have any questions or feedback.