bert-finetuned-ner-optuna

This model is a fine-tuned version of bert-base-cased on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.6380
• Precision: 0.2930
• Recall: 0.2959
• F1: 0.2944
• Accuracy: 0.8208

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: 1.0447410202448447e-06
• train_batch_size: 16
• eval_batch_size: 8
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1.634485341029132e-09
• lr_scheduler_type: linear
• num_epochs: 10

Training results

Training Loss	Epoch	Step	Validation Loss	Precision	Recall	F1	Accuracy
No log	1.0	62	1.7624	0.0382	0.0134	0.0199	0.7006
No log	2.0	124	1.1267	0.0	0.0	0.0	0.7166
No log	3.0	186	0.9232	0.0155	0.0099	0.0121	0.7438
No log	4.0	248	0.8140	0.1012	0.0713	0.0837	0.7747
No log	5.0	310	0.7416	0.2075	0.1907	0.1987	0.7941
No log	6.0	372	0.6991	0.2582	0.2620	0.2601	0.8054
No log	7.0	434	0.6692	0.2809	0.2825	0.2817	0.8138
No log	8.0	496	0.6484	0.2905	0.2888	0.2897	0.8201
No log	9.0	558	0.6425	0.2861	0.2895	0.2878	0.8191
No log	10.0	620	0.6380	0.2930	0.2959	0.2944	0.8208

Framework versions

• Transformers 4.34.1
• Pytorch 2.1.0
• Datasets 2.14.6
• Tokenizers 0.14.1