final-ft__roberta-base-bne__70k-ultrasounds-ner

This model is a fine-tuned version of manucos/final-ft__roberta-base-bne__70k-ultrasounds on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.3223
• Precision: 0.8759
• Recall: 0.9096
• F1: 0.8924
• Accuracy: 0.9437

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: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 15

Training results

Training Loss	Epoch	Step	Validation Loss	Precision	Recall	F1	Accuracy
No log	1.0	280	0.2838	0.8408	0.8607	0.8507	0.9286
0.3392	2.0	560	0.2263	0.8261	0.88	0.8522	0.9345
0.3392	3.0	840	0.2402	0.8528	0.8667	0.8597	0.9357
0.0876	4.0	1120	0.2296	0.8425	0.9037	0.8721	0.9426
0.0876	5.0	1400	0.2738	0.8559	0.8978	0.8764	0.9416
0.0391	6.0	1680	0.2462	0.8517	0.9022	0.8763	0.9412
0.0391	7.0	1960	0.2798	0.8737	0.9022	0.8878	0.9428
0.0196	8.0	2240	0.2841	0.8596	0.8978	0.8783	0.9439
0.0159	9.0	2520	0.3217	0.8524	0.9067	0.8787	0.9426
0.0159	10.0	2800	0.3079	0.8771	0.9096	0.8931	0.9446
0.0118	11.0	3080	0.3166	0.8694	0.8978	0.8834	0.9426
0.0118	12.0	3360	0.3160	0.8711	0.9111	0.8907	0.9451
0.0099	13.0	3640	0.3178	0.8759	0.9096	0.8924	0.9437
0.0099	14.0	3920	0.3214	0.8759	0.9096	0.8924	0.9437
0.0086	15.0	4200	0.3223	0.8759	0.9096	0.8924	0.9437

Framework versions

• Transformers 4.41.1
• Pytorch 2.3.0+cu121
• Datasets 2.19.1
• Tokenizers 0.19.1