06-12-15-37

This model is a fine-tuned version of dbmdz/bert-base-turkish-cased on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.2773
• Hate Precision: 0.7778
• Hate Recall: 0.5385
• Hate F1: 0.6364
• Sexual Precision: 0.0
• Sexual Recall: 0.0
• Sexual F1: 0.0
• Threat Precision: 0.0
• Threat Recall: 0.0
• Threat F1: 0.0
• Neutral Precision: 0.8378
• Neutral Recall: 0.9688
• Neutral F1: 0.8986
• Macro Precision: 0.4039
• Macro Recall: 0.3768
• Macro F1: 0.3837

Model description

More information needed

Intended uses & limitations

More information needed

Training and evaluation data

More information needed

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: Use OptimizerNames.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: 10

Training results

Training Loss	Epoch	Step	Validation Loss	Hate Precision	Hate Recall	Hate F1	Sexual Precision	Sexual Recall	Sexual F1	Threat Precision	Threat Recall	Threat F1	Neutral Precision	Neutral Recall	Neutral F1	Macro Precision	Macro Recall	Macro F1
0.6107	0.8333	10	0.4510	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.7111	1.0	0.8312	0.1778	0.25	0.2078
0.4163	1.6667	20	0.3613	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.7111	1.0	0.8312	0.1778	0.25	0.2078
0.3561	2.5	30	0.3459	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.7111	1.0	0.8312	0.1778	0.25	0.2078
0.3446	3.3333	40	0.3395	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.7111	1.0	0.8312	0.1778	0.25	0.2078
0.3002	4.1667	50	0.3270	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.7143	0.9375	0.8108	0.1786	0.2344	0.2027
0.2912	5.0	60	0.3357	0.75	0.2308	0.3529	0.0	0.0	0.0	0.0	0.0	0.0	0.7381	0.9688	0.8378	0.3720	0.2999	0.2977
0.2355	5.8333	70	0.3359	0.75	0.2308	0.3529	0.0	0.0	0.0	0.0	0.0	0.0	0.7561	0.9688	0.8493	0.3765	0.2999	0.3006
0.2017	6.6667	80	0.2989	0.6667	0.4615	0.5455	0.0	0.0	0.0	0.0	0.0	0.0	0.8056	0.9062	0.8529	0.3681	0.3419	0.3496
0.1988	7.5	90	0.2904	0.7	0.5385	0.6087	0.0	0.0	0.0	0.0	0.0	0.0	0.8286	0.9062	0.8657	0.3821	0.3612	0.3686
0.1476	8.3333	100	0.3351	0.6667	0.1538	0.25	0.0	0.0	0.0	0.0	0.0	0.0	0.7381	0.9688	0.8378	0.3512	0.2806	0.2720
0.1846	9.1667	110	0.2774	0.7778	0.5385	0.6364	0.0	0.0	0.0	0.0	0.0	0.0	0.8333	0.9375	0.8824	0.4028	0.3690	0.3797
0.1347	10.0	120	0.2773	0.7778	0.5385	0.6364	0.0	0.0	0.0	0.0	0.0	0.0	0.8378	0.9688	0.8986	0.4039	0.3768	0.3837

Framework versions

• Transformers 4.47.0
• Pytorch 2.5.1+cu124
• Datasets 3.1.0
• Tokenizers 0.21.0