bert-base-turkish-cased-stopword-product-names-classification

This model is a fine-tuned version of dbmdz/bert-base-turkish-cased on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.3536
• Accuracy: 0.948

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: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 30

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
No log	1.0	313	2.2729	0.5744
3.0971	2.0	626	1.0825	0.8184
3.0971	3.0	939	0.7240	0.8616
0.9848	4.0	1252	0.5729	0.8776
0.4744	5.0	1565	0.4712	0.8944
0.4744	6.0	1878	0.4396	0.9024
0.2716	7.0	2191	0.3794	0.9176
0.1782	8.0	2504	0.3714	0.932
0.1782	9.0	2817	0.3588	0.928
0.1104	10.0	3130	0.3214	0.9384
0.1104	11.0	3443	0.3391	0.9376
0.0843	12.0	3756	0.3376	0.9408
0.0616	13.0	4069	0.3470	0.9384
0.0616	14.0	4382	0.3525	0.9336
0.0384	15.0	4695	0.3335	0.9368
0.033	16.0	5008	0.3334	0.9424
0.033	17.0	5321	0.3490	0.94
0.0246	18.0	5634	0.3260	0.944
0.0246	19.0	5947	0.3482	0.9408
0.0173	20.0	6260	0.3482	0.9456
0.0138	21.0	6573	0.3446	0.9448
0.0138	22.0	6886	0.3578	0.944
0.008	23.0	7199	0.3415	0.9464
0.007	24.0	7512	0.3536	0.948
0.007	25.0	7825	0.3667	0.948
0.0069	26.0	8138	0.3605	0.9464
0.0069	27.0	8451	0.3652	0.9456
0.0053	28.0	8764	0.3650	0.9472
0.0043	29.0	9077	0.3623	0.9472
0.0043	30.0	9390	0.3635	0.9472

Framework versions

• Transformers 4.33.0
• Pytorch 2.0.0
• Datasets 2.1.0
• Tokenizers 0.13.3