movie-review-classifier

movie-review-classifier

This model classifies (text) movie reviews as either a 1 (i.e., thumbs-up) or a 0 (i.e., a thumbs-down).

Model description

This model is a version of distilbert-base-uncased that was fine-tuned on the IMDB movie-review dataset. It achieves the following results on the evaluation set:

• Loss: 0.2743
• F1: 0.9327

Intended uses & limitations

Training this model was completed as part of a project from a data science bootcamp. It is intended to be used perhaps by students and/or hobbyists.

Training and evaluation data

This model was trained on the IMDB movie-review dataset, a set of highly polarized (i.e., clearly positive or negative) movie reviews. The dataset contains 25k labelled train samples, 25k labelled test samples, and 50k unlabelled samples.

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: 3
• weight_decay: 0.1

Training results

Framework versions

• Transformers 4.42.4
• Pytorch 2.3.1+cu121
• Datasets 2.20.0
• Tokenizers 0.19.1