roberta-large-NER_keyword_oknashar

This model was trained from scratch on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.2412
• Precision: 0.7773
• Recall: 0.7947
• F1: 0.7859
• Accuracy: 0.9732

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: 4
• eval_batch_size: 8
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 4
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Precision	Recall	F1	Accuracy
0.007	1	2714	0.2217	0.7532	0.7489	0.7510	0.9705
0.0056	2	5428	0.2128	0.7281	0.7752	0.7509	0.9701
0.008	3	8142	0.2351	0.7650	0.7821	0.7735	0.9713
0.008	4	10856	0.2183	0.7579	0.7792	0.7684	0.9716
0.0061	5	2714	0.2430	0.7534	0.7219	0.7373	0.9692
0.0061	6	5428	0.2506	0.7570	0.7752	0.7660	0.9716
0.013	7	8142	0.2322	0.7583	0.7557	0.7570	0.9708
0.0085	8	10856	0.2129	0.7616	0.7838	0.7725	0.9721
0.002	9	13570	0.2259	0.7683	0.7815	0.7749	0.9724
0.0071	10	2714	0.2433	0.7619	0.7506	0.7562	0.9707
0.0071	11	5428	0.2490	0.7456	0.7615	0.7535	0.9699
0.0035	12	8142	0.2483	0.7788	0.7873	0.7830	0.9727
0.002	13	10856	0.2412	0.7773	0.7947	0.7859	0.9732

Framework versions

• Transformers 4.35.1
• Pytorch 2.1.0+cu118
• Datasets 2.14.6
• Tokenizers 0.14.1