modernBert-base_v2

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

• Loss: 0.7185
• Accuracy: 0.9116
• Precision Macro: 0.8041
• Recall Macro: 0.7362
• F1 Macro: 0.7592
• F1 Weighted: 0.9065

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: 3e-05
• train_batch_size: 64
• eval_batch_size: 64
• seed: 42
• gradient_accumulation_steps: 2
• total_train_batch_size: 128
• 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
• num_epochs: 20
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	Precision Macro	Recall Macro	F1 Macro	F1 Weighted
1.2139	1.0	90	0.5024	0.8073	0.8182	0.5993	0.6061	0.7934
0.6774	2.0	180	0.2870	0.9033	0.8421	0.7140	0.7451	0.8960
0.4571	3.0	270	0.3474	0.8920	0.8074	0.6669	0.6824	0.8802
0.2925	4.0	360	0.3089	0.9065	0.8778	0.7074	0.7413	0.8977
0.1725	5.0	450	0.3611	0.8958	0.7729	0.7574	0.7646	0.8946
0.0977	6.0	540	0.4743	0.9090	0.8405	0.7388	0.7695	0.9036
0.0576	7.0	630	0.6044	0.8743	0.7234	0.8019	0.7413	0.8878
0.0338	8.0	720	0.6118	0.9040	0.7756	0.7506	0.7615	0.9019
0.016	9.0	810	0.6754	0.9071	0.8334	0.7379	0.7670	0.9019
0.0113	10.0	900	0.6732	0.9065	0.7898	0.7606	0.7733	0.9044
0.0065	11.0	990	0.7871	0.9046	0.8046	0.7277	0.7519	0.8992
0.0037	12.0	1080	0.7134	0.9109	0.7989	0.7147	0.7386	0.9038
0.0022	13.0	1170	0.7784	0.9015	0.7765	0.7383	0.7529	0.8982
0.0013	14.0	1260	0.7176	0.9109	0.7832	0.7486	0.7625	0.9079
0.0011	15.0	1350	0.7681	0.9059	0.7920	0.7371	0.7565	0.9017
0.0001	16.0	1440	0.7170	0.9071	0.7833	0.7282	0.7479	0.9024
0.0007	17.0	1530	0.7219	0.9109	0.8022	0.7442	0.7652	0.9068
0.0003	18.0	1620	0.7379	0.9103	0.7950	0.7398	0.7596	0.9060
0.0006	19.0	1710	0.7198	0.9116	0.8074	0.7404	0.7635	0.9068
0.0004	20.0	1800	0.7185	0.9116	0.8041	0.7362	0.7592	0.9065

Framework versions

• Transformers 4.55.0
• Pytorch 2.7.0+cu126
• Datasets 4.0.0
• Tokenizers 0.21.4