bert-base-italian-xxl-cased-sentence-splitter

This model is a fine-tuned version of dbmdz/bert-base-italian-xxl-cased on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.0021
• F1: 0.9938

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: 8
• eval_batch_size: 8
• seed: 42
• 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
• num_epochs: 30

Training results

Training Loss	Epoch	Step	Validation Loss	F1
No log	1.0	49	0.0036	0.9831
No log	2.0	98	0.0021	0.9907
No log	3.0	147	0.0030	0.9861
No log	4.0	196	0.0026	0.9907
No log	5.0	245	0.0018	0.9938
No log	6.0	294	0.0020	0.9938
No log	7.0	343	0.0040	0.9861
No log	8.0	392	0.0023	0.9922
No log	9.0	441	0.0024	0.9922
No log	10.0	490	0.0062	0.9922
0.0139	11.0	539	0.0045	0.9891
0.0139	12.0	588	0.0019	0.9922
0.0139	13.0	637	0.0021	0.9938
0.0139	14.0	686	0.0024	0.9938
0.0139	15.0	735	0.0120	0.9891
0.0139	16.0	784	0.0074	0.9907
0.0139	17.0	833	0.0019	0.9938
0.0139	18.0	882	0.0019	0.9938
0.0139	19.0	931	0.0024	0.9922
0.0139	20.0	980	0.0021	0.9922
0.0002	21.0	1029	0.0021	0.9922
0.0002	22.0	1078	0.0021	0.9938
0.0002	23.0	1127	0.0022	0.9922
0.0002	24.0	1176	0.0020	0.9938
0.0002	25.0	1225	0.0022	0.9938
0.0002	26.0	1274	0.0021	0.9938
0.0002	27.0	1323	0.0022	0.9938
0.0002	28.0	1372	0.0021	0.9938
0.0002	29.0	1421	0.0021	0.9938
0.0002	30.0	1470	0.0021	0.9938

Framework versions

• Transformers 4.55.2
• Pytorch 2.8.0+cu128
• Datasets 3.6.0
• Tokenizers 0.21.4