bert-finetuned-ner-05-no-loss-on-padding

This model is a fine-tuned version of FacebookAI/roberta-base on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.0422
• Precision: 0.9445
• Recall: 0.9567
• F1: 0.9505
• Accuracy: 0.9891

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: 32
• eval_batch_size: 32
• 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: 3

Training results

Training Loss	Epoch	Step	Validation Loss	Precision	Recall	F1	Accuracy
0.2251	0.4556	200	0.0758	0.8727	0.9087	0.8903	0.9783
0.0724	0.9112	400	0.0542	0.9182	0.9360	0.9270	0.9841
0.0516	1.3667	600	0.0486	0.9308	0.9469	0.9388	0.9869
0.0428	1.8223	800	0.0437	0.9400	0.9518	0.9459	0.9883
0.0306	2.2779	1000	0.0432	0.9415	0.9567	0.9490	0.9887
0.0267	2.7335	1200	0.0427	0.9420	0.9555	0.9487	0.9890
0.0267	3.0	1317	0.0422	0.9445	0.9567	0.9505	0.9891

Framework versions

• Transformers 5.7.0
• Pytorch 2.10.0+cu128
• Datasets 4.0.0
• Tokenizers 0.22.2