ModernBERT-Turkish-Sentiment-Classifier

This model is a fine-tuned version of artiwise-ai/modernbert-base-tr-uncased on a Sentiment Analysis dataset winvoker/turkish-sentiment-analysis-dataset. It achieves the following results on the evaluation set:

• Loss: 0.280550
• Accuracy: 0.956000
• Precision: 0.956370
• Recall: 0.956000
• F1: 0.956175

Training and evaluation data

The data cannot be shared.

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

• learning_rate: 5e-05
• train_batch_size: 16
• eval_batch_size: 16
• seed: 42
• optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• num_epochs: 3

Training results

Epoch	Training Loss	Validation Loss	Accuracy	Precision	Recall	F1 Score
1	0.161400	0.163402	0.945000	0.945220	0.945000	0.945107
2	0.034400	0.286277	0.942000	0.949040	0.942000	0.944231
3	0.003100	0.280550	0.956000	0.956370	0.956000	0.956170

Framework versions

• Transformers 4.52.2
• Pytorch 2.6.0+cu124
• Datasets 2.14.4
• Tokenizers 0.21.1