NER-challenge

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

• Loss: 0.2979
• Precision: 0.8
• Recall: 0.5714
• F1: 0.6667
• Accuracy: 0.9048

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: 15

Training results

Training Loss	Epoch	Step	Validation Loss	Precision	Recall	F1	Accuracy
No log	1.0	2	1.1604	0.1351	0.0893	0.1075	0.7415
No log	2.0	4	0.9221	0.0	0.0	0.0	0.7959
No log	3.0	6	0.7388	0.0	0.0	0.0	0.7959
No log	4.0	8	0.6175	0.0	0.0	0.0	0.7959
No log	5.0	10	0.5558	0.0	0.0	0.0	0.7959
No log	6.0	12	0.5125	0.0	0.0	0.0	0.7959
No log	7.0	14	0.4720	0.0	0.0	0.0	0.7959
No log	8.0	16	0.4339	0.0	0.0	0.0	0.7959
No log	9.0	18	0.3996	0.375	0.1071	0.1667	0.8231
No log	10.0	20	0.3702	0.6538	0.3036	0.4146	0.8571
No log	11.0	22	0.3457	0.7647	0.4643	0.5778	0.8844
No log	12.0	24	0.3261	0.8	0.5714	0.6667	0.9048
No log	13.0	26	0.3115	0.8	0.5714	0.6667	0.9048
No log	14.0	28	0.3020	0.8	0.5714	0.6667	0.9048
No log	15.0	30	0.2979	0.8	0.5714	0.6667	0.9048

Framework versions

• Transformers 4.40.1
• Pytorch 2.2.1+cu121
• Datasets 2.19.1
• Tokenizers 0.19.1