valueeval24-modern-bert

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: 0.1613
• F1: 0.3178
• Roc Auc: 0.6190
• Accuracy: 0.1954

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-06
• train_batch_size: 8
• eval_batch_size: 8
• seed: 2024
• 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_ratio: 0.01
• num_epochs: 20
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	F1	Roc Auc	Accuracy
0.1463	1.0	2883	0.1052	0.1633	0.5464	0.0854
0.1003	2.0	5766	0.0995	0.2146	0.5640	0.1188
0.0907	3.0	8649	0.0981	0.2777	0.5899	0.1662
0.0806	4.0	11532	0.1001	0.3038	0.6035	0.1804
0.0685	5.0	14415	0.1048	0.3099	0.6094	0.1914
0.0549	6.0	17298	0.1104	0.3209	0.6177	0.1968
0.0412	7.0	20181	0.1158	0.3197	0.6198	0.1934
0.0285	8.0	23064	0.1232	0.3226	0.6210	0.1974
0.0184	9.0	25947	0.1312	0.3157	0.6186	0.1943
0.0114	10.0	28830	0.1381	0.3176	0.6192	0.1951
0.0071	11.0	31713	0.1463	0.3216	0.6216	0.1972
0.0047	12.0	34596	0.1542	0.3153	0.6168	0.1959
0.0032	13.0	37479	0.1613	0.3178	0.6190	0.1954

Framework versions

• Transformers 4.53.0
• Pytorch 2.5.1+cu121
• Datasets 3.6.0
• Tokenizers 0.21.2