bert2

This model is a fine-tuned version of xlm-roberta-base on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.0826
• Accuracy: 0.9868
• F1: 0.9869
• Precision: 0.9874
• Recall: 0.9868

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: 5e-05
• train_batch_size: 16
• eval_batch_size: 16
• seed: 42
• optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 500
• num_epochs: 20

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	F1	Precision	Recall
1.0962	1.0	38	1.0876	0.5364	0.4557	0.4007	0.5364
1.0434	2.0	76	0.9229	0.5894	0.4961	0.4283	0.5894
0.5945	3.0	114	0.3387	0.9205	0.9208	0.9318	0.9205
0.2215	4.0	152	0.2386	0.9272	0.9274	0.9407	0.9272
0.0876	5.0	190	0.1032	0.9735	0.9735	0.9736	0.9735
0.0236	6.0	228	0.3850	0.9404	0.9406	0.9498	0.9404
0.0891	7.0	266	0.0983	0.9801	0.9802	0.9813	0.9801
0.0606	8.0	304	0.3116	0.9470	0.9472	0.9546	0.9470
0.0006	9.0	342	0.3678	0.9470	0.9472	0.9546	0.9470
0.0821	10.0	380	0.0105	1.0	1.0	1.0	1.0
0.0057	11.0	418	0.3436	0.9603	0.9604	0.9647	0.9603
0.1437	12.0	456	0.1034	0.9735	0.9736	0.9741	0.9735
0.0615	13.0	494	0.4131	0.9404	0.9411	0.9498	0.9404
0.0697	14.0	532	0.1664	0.9735	0.9739	0.9759	0.9735
0.0242	15.0	570	0.0595	0.9934	0.9934	0.9935	0.9934
0.0002	16.0	608	0.1706	0.9735	0.9736	0.9755	0.9735
0.0002	17.0	646	0.0598	0.9934	0.9934	0.9935	0.9934
0.0002	18.0	684	0.0776	0.9868	0.9869	0.9874	0.9868
0.0002	19.0	722	0.0805	0.9868	0.9869	0.9874	0.9868
0.0002	20.0	760	0.0826	0.9868	0.9869	0.9874	0.9868

Framework versions

• Transformers 4.50.0.dev0
• Pytorch 2.5.1+cu121
• Datasets 3.4.1
• Tokenizers 0.21.0