finetuned-test-1

This model is a fine-tuned version of bert-base-uncased on the conll2003 dataset. It achieves the following results on the evaluation set:

• Loss: 1.8192

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: 16
• eval_batch_size: 16
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 20
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss
2.8219	1.0	30	2.3343
2.4148	2.0	60	2.2010
2.3236	3.0	90	2.1442
2.2231	4.0	120	2.1651
2.2171	5.0	150	2.0614
2.127	6.0	180	2.0405
2.0748	7.0	210	2.0092
2.0511	8.0	240	1.9798
2.0097	9.0	270	1.8662
1.9969	10.0	300	1.9257
2.0006	11.0	330	1.9386
1.9273	12.0	360	1.9357
1.9177	13.0	390	1.8983
1.9128	14.0	420	1.8990
1.8979	15.0	450	1.9037
1.8721	16.0	480	1.8440
1.8998	17.0	510	1.8404
1.8862	18.0	540	1.9193
1.9133	19.0	570	1.8494
1.8799	20.0	600	1.8192

Framework versions

• Transformers 4.20.1
• Pytorch 1.11.0+cu113
• Datasets 2.3.2
• Tokenizers 0.12.1