eraydikyologlu/bert_ayt_matematik

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:

• Train Loss: 0.0544
• Train Accuracy: 0.9874
• Validation Loss: 0.0221
• Validation Accuracy: 0.9969
• Epoch: 24

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:

• optimizer: {'name': 'Adam', 'weight_decay': None, 'clipnorm': None, 'global_clipnorm': None, 'clipvalue': None, 'use_ema': False, 'ema_momentum': 0.99, 'ema_overwrite_frequency': None, 'jit_compile': True, 'is_legacy_optimizer': False, 'learning_rate': {'module': 'transformers.optimization_tf', 'class_name': 'WarmUp', 'config': {'initial_learning_rate': 2e-05, 'decay_schedule_fn': {'module': 'keras.optimizers.schedules', 'class_name': 'PolynomialDecay', 'config': {'initial_learning_rate': 2e-05, 'decay_steps': 11295, 'end_learning_rate': 0.0, 'power': 1.0, 'cycle': False, 'name': None}, 'registered_name': None}, 'warmup_steps': 1255, 'power': 1.0, 'name': None}, 'registered_name': 'WarmUp'}, 'beta_1': 0.9, 'beta_2': 0.999, 'epsilon': 1e-08, 'amsgrad': False}
• training_precision: float32

Training results

Train Loss	Train Accuracy	Validation Loss	Validation Accuracy	Epoch
3.6700	0.1020	2.8781	0.3262	0
2.2363	0.4466	1.5455	0.5706	1
1.4585	0.5886	1.1438	0.6481	2
1.1586	0.6523	0.9413	0.7075	3
0.9850	0.6957	0.7765	0.7431	4
0.8494	0.7339	0.6264	0.7944	5
0.7308	0.7697	0.5343	0.8250	6
0.6252	0.8038	0.4364	0.8644	7
0.5335	0.8339	0.3167	0.9087	8
0.4434	0.8632	0.2438	0.9362	9
0.3789	0.8856	0.1952	0.9475	10
0.3162	0.9034	0.1544	0.9631	11
0.2662	0.9213	0.1115	0.9744	12
0.2234	0.9362	0.0941	0.9756	13
0.1857	0.9465	0.0782	0.9812	14
0.1608	0.9546	0.0554	0.9887	15
0.1390	0.9615	0.0569	0.9881	16
0.1222	0.9668	0.0436	0.9887	17
0.1055	0.9714	0.0357	0.9931	18
0.0930	0.9756	0.0356	0.9912	19
0.0817	0.9793	0.0304	0.9950	20
0.0747	0.9814	0.0266	0.9944	21
0.0660	0.9838	0.0244	0.9969	22
0.0586	0.9863	0.0223	0.9975	23
0.0544	0.9874	0.0221	0.9969	24

Framework versions

• Transformers 4.52.4
• TensorFlow 2.18.0
• Datasets 2.14.4
• Tokenizers 0.21.1