ner_bert_model

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

• Loss: 0.4145
• Precision: 0.8235
• Recall: 0.9333
• F1: 0.8750
• Accuracy: 0.9096

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: 2
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 20

Training results

Training Loss	Epoch	Step	Validation Loss	Precision	Recall	F1	Accuracy
No log	1.0	7	1.7796	0.0	0.0	0.0	0.4294
No log	2.0	14	1.4530	0.5	0.2667	0.3478	0.5650
No log	3.0	21	1.1854	0.5510	0.36	0.4355	0.6384
No log	4.0	28	0.9850	0.6667	0.5867	0.6241	0.7345
No log	5.0	35	0.8189	0.6622	0.6533	0.6577	0.7797
No log	6.0	42	0.7194	0.6914	0.7467	0.7179	0.8192
No log	7.0	49	0.6126	0.7262	0.8133	0.7673	0.8588
No log	8.0	56	0.5760	0.75	0.88	0.8098	0.8701
No log	9.0	63	0.4819	0.8	0.9067	0.8500	0.8927
No log	10.0	70	0.4610	0.7907	0.9067	0.8447	0.8983
No log	11.0	77	0.4471	0.8	0.9067	0.8500	0.8927
No log	12.0	84	0.4203	0.7931	0.92	0.8519	0.9040
No log	13.0	91	0.4281	0.8256	0.9467	0.8820	0.9153
No log	14.0	98	0.3913	0.8256	0.9467	0.8820	0.9153
No log	15.0	105	0.3966	0.8235	0.9333	0.8750	0.9096
No log	16.0	112	0.4033	0.8235	0.9333	0.8750	0.9096
No log	17.0	119	0.4149	0.8140	0.9333	0.8696	0.9040
No log	18.0	126	0.4150	0.8140	0.9333	0.8696	0.9040
No log	19.0	133	0.4122	0.8235	0.9333	0.8750	0.9096
No log	20.0	140	0.4145	0.8235	0.9333	0.8750	0.9096

Framework versions

• Transformers 4.39.1
• Pytorch 2.2.1+cu121
• Datasets 2.18.0
• Tokenizers 0.15.2