distilbert-base-uncased-finetuned-emotion-movies-186k

This model is a fine-tuned version of distilbert-base-uncased on a 186k movie reviews/emotions self-collected dataset from 1150 movies from TMDB. It achieves the following results on the evaluation set:

• Loss: 0.3572
• Accuracy: 0.8635
• F1: 0.8637

Model description

The model classifies into the following emotions:

• 'LABEL_0': 'sadness'
• 'LABEL_1': 'joy'
• 'LABEL_2': 'love'
• 'LABEL_3': 'anger'
• 'LABEL_4': 'fear'
• 'LABEL_5': 'surprise'

Intended uses & limitations

Academic

Training and evaluation data

The model was trained with a dataset (186k rows) of movies reviews/emotions from 1150 movies from TMDB, taking 20% for testing.

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

• learning_rate: 2e-05
• train_batch_size: 32
• eval_batch_size: 32
• 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	F1
0.4956	1.0	5828	0.3770	0.8531	0.8513
0.3035	2.0	11656	0.3572	0.8635	0.8637

Framework versions

• Transformers 4.31.0
• Pytorch 2.0.1+cu118
• Datasets 2.14.4
• Tokenizers 0.13.3