results

This model is a fine-tuned version of distilbert-base-uncased on an Emotion Dataset.

Model description

This model fine-tunes DistilBERT for emotion classification. It can detect emotions in language and then classify them into: sadness, joy, love, anger, fear, surprise.

Intended uses & limitations

Used to explain the inner emotions of simple sentences. This model may lack contextual reasoning ability and cannot understand connecting words such as transitions.

Training and evaluation data

• Training Dataset: dair-ai/emotion (16,000 examples)
• Validation set: 2,000 examples
• Test set: 2,000 examples
• Validation Accuracy:
  • epoch1:0.9065
  • epoch2:0.9345
  • epoch3:0.93
  • epoch4:0.942
  • epoch5:0.94
• Test Accuracy: 0.942
• Training Time: 2:02:44

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

• learning_rate: 5e-05
• train_batch_size: 256
• eval_batch_size: 256
• 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: 5

Framework versions

• Transformers 4.46.2
• Pytorch 2.5.1+cpu
• Datasets 3.1.0
• Tokenizers 0.20.3