roberta-base-finetuned-abbr-finetuned-ner

This model is a fine-tuned version of surrey-nlp/roberta-base-finetuned-abbr on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.1753
• Precision: 0.9674
• Recall: 0.9681
• F1: 0.9678
• Accuracy: 0.9618

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-06
• train_batch_size: 16
• eval_batch_size: 4
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 64
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 10
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Precision	Recall	F1	Accuracy
No log	0.6	10	0.9865	0.7657	0.7982	0.7816	0.7573
No log	1.19	20	0.7172	0.8161	0.8566	0.8359	0.8204
No log	1.79	30	0.5382	0.8437	0.8759	0.8595	0.8478
No log	2.39	40	0.4196	0.8713	0.8938	0.8824	0.8733
No log	2.99	50	0.3485	0.8965	0.9112	0.9038	0.8979
No log	3.58	60	0.3031	0.9241	0.9325	0.9283	0.9218
No log	4.18	70	0.2688	0.9459	0.9496	0.9477	0.9411
No log	4.78	80	0.2434	0.9531	0.9559	0.9545	0.9481
No log	5.37	90	0.2235	0.9605	0.9623	0.9614	0.9555
No log	5.97	100	0.2078	0.9612	0.9623	0.9618	0.9559
No log	6.57	110	0.1966	0.9637	0.9647	0.9642	0.9580
No log	7.16	120	0.1879	0.9646	0.9655	0.9651	0.9591
No log	7.76	130	0.1821	0.9664	0.9671	0.9667	0.9608
No log	8.36	140	0.1782	0.9669	0.9676	0.9673	0.9613
No log	8.96	150	0.1760	0.9674	0.9683	0.9679	0.9618
No log	9.55	160	0.1753	0.9674	0.9681	0.9678	0.9618

Framework versions

• Transformers 4.39.3
• Pytorch 2.2.2+cu121
• Datasets 2.19.0
• Tokenizers 0.15.2