modernbert-binary-disfluency-finetuned

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

• Loss: 0.0533
• Accuracy: 0.9785
• Precision: 0.9190
• Recall: 0.9190
• F1: 0.9190
• Specificity: 0.9876
• True Positives: 703
• False Positives: 62
• True Negatives: 4950
• False Negatives: 62

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: 32
• eval_batch_size: 8
• seed: 42
• optimizer: Use OptimizerNames.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: 0.1
• num_epochs: 5
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	Precision	Recall	F1	Specificity	True Positives	False Positives	True Negatives	False Negatives
No log	0.5747	50	0.0297	0.9846	0.9340	0.9462	0.9400	0.9902	580	41	4154	33
No log	1.1494	100	0.0348	0.9846	0.9354	0.9445	0.9399	0.9905	579	40	4155	34
No log	1.7241	150	0.0303	0.9848	0.9383	0.9429	0.9406	0.9909	578	38	4157	35
No log	2.2989	200	0.0249	0.9827	0.9102	0.9592	0.9341	0.9862	588	58	4137	25
No log	2.8736	250	0.0303	0.9832	0.9196	0.9511	0.9350	0.9878	583	51	4144	30
No log	3.4483	300	0.0276	0.9819	0.9109	0.9511	0.9306	0.9864	583	57	4138	30
No log	4.0230	350	0.0283	0.9817	0.9082	0.9527	0.9299	0.9859	584	59	4136	29
No log	4.5977	400	0.0294	0.9825	0.9152	0.9511	0.9328	0.9871	583	54	4141	30

Framework versions

• Transformers 4.48.3
• Pytorch 2.6.0+cu124
• Datasets 3.4.0
• Tokenizers 0.21.0