emotion-advance-classifier
This model is a fine-tuned version of microsoft/MiniLM-L12-H384-uncased on the Emotion Dataset 2.0[https://huggingface.co/datasets/Keyurjotaniya007/datasets-emotion-2.0]. It achieves the following results on the evaluation set:
• Loss: 0.22426407039165497
• Accuracy: 0.919
• F1: 0.9197963043356645
This model is trained and evaluated using 'emotion' dataset.
Emotion Dataset 2.0 (Balanced & Augmented)
This is a balanced version of the famous dair-ai/emotion dataset, created using NLP data augmentation techniques to handle class imbalance and improve model robustness.
Emotion Dataset 2.0(Class Distribution)
| Emotion | Count |
|---|---|
| sadness | 5362 |
| joy | 5362 |
| love | 5362 |
| anger | 5362 |
| fear | 5362 |
| surprise | 5362 |
Training hyperparameters
The following hyperparameters were used during training:
• learning_rate: 2e-05
• train_batch_size: 16
• eval_batch_size: 16
• seed: 42
• weight_decay: 0.01
• lr_scheduler_type: linear
• warmup_ratio: 0.1
• num_epochs: 3
Framework versions
• Transformers 4.31.0
• Pytorch 2.0.1+cu118
• Datasets 2.14.4
• Tokenizers 0.13.3
- Downloads last month
- 22