modernbert-soft-ste-sentence-pool-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: 0.0591
• Micro F1: 0.5528
• Micro Precision: 0.8430
• Micro Recall: 0.4113
• F1 Active Voice: 0.4286
• F1 Clear Reference: 0.0645
• F1 Clear Sentences: 0.5
• F1 Condition First: 0.2222
• F1 Direct Report: 0.5000
• F1 Direct Verbs: 0.6
• F1 Logical Order: 0.0
• F1 No Contractions: 0.9455
• F1 No Semicolon: 1.0
• F1 One Idea: 0.4865
• F1 Plain Language: 0.5714
• F1 Simple Verbs: 0.9

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: 2e-05
• train_batch_size: 4
• eval_batch_size: 4
• seed: 100
• optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED 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: 10

Training results

Training Loss	Epoch	Step	Validation Loss	Micro F1	Micro Precision	Micro Recall	F1 Active Voice	F1 Clear Reference	F1 Clear Sentences	F1 Condition First	F1 Direct Report	F1 Direct Verbs	F1 Logical Order	F1 No Contractions	F1 No Semicolon	F1 One Idea	F1 Plain Language	F1 Simple Verbs
0.0978	1.0	366	0.1039	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0
0.0693	2.0	732	0.0738	0.3410	0.9123	0.2097	0.2051	0.0	0.0	0.0	0.0	0.1333	0.0	0.8519	0.9412	0.0	0.4516	0.1667
0.0458	3.0	1098	0.0624	0.4510	0.8539	0.3065	0.1026	0.0	0.4706	0.0	0.1667	0.56	0.0	0.8929	1.0	0.0690	0.5294	0.9
0.0379	4.0	1464	0.0561	0.5639	0.7006	0.4718	0.4857	0.0	0.375	0.2222	0.4286	0.6667	0.0	0.9667	1.0	0.5085	0.5714	0.9
0.0219	5.0	1830	0.0550	0.5485	0.6890	0.4556	0.5	0.2353	0.0	0.08	0.5000	0.7273	0.0	0.9492	1.0	0.5366	0.5294	0.9
0.0096	6.0	2196	0.0596	0.5455	0.8095	0.4113	0.4262	0.0	0.4	0.3125	0.5000	0.4444	0.0	0.9455	1.0	0.6	0.5294	0.8421
0.0063	7.0	2562	0.0591	0.5528	0.8430	0.4113	0.4286	0.0645	0.5	0.2222	0.5000	0.6	0.0	0.9455	1.0	0.4865	0.5714	0.9

Framework versions

• Transformers 4.57.6
• Pytorch 2.13.0+cu130
• Tokenizers 0.22.2