CeLLaTe-ner-2class-pubmedbert-tapt-tokenizer-adapted-spanmask-combData-lr_4.636

This model is a fine-tuned version of Mardiyyah/CeLLaTe-tapt-pubmedbert-tokenizer-adapted-spanmask-combinedData on the OTAR3088/CeLLaTe-ner-2class-iob_final dataset. It achieves the following results on the evaluation set:

• Loss: 0.0994
• Precision: 0.7792
• Recall: 0.7619
• Micro F1: 0.7704
• Weighted F1: 0.7707
• Macro F1: 0.7778
• Accuracy: 0.9844

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: 4.6356202949590504e-05
• train_batch_size: 32
• eval_batch_size: 16
• seed: 17
• gradient_accumulation_steps: 2
• total_train_batch_size: 64
• optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: cosine
• lr_scheduler_warmup_ratio: 0.01
• num_epochs: 20
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Precision	Recall	Micro F1	Weighted F1	Macro F1	Accuracy
0.3296	1.0	132	0.0846	0.4061	0.2742	0.3274	0.3219	0.2575	0.9696
0.0449	2.0	264	0.0606	0.6839	0.7300	0.7062	0.7081	0.7237	0.9822
0.0257	3.0	396	0.0622	0.7778	0.6861	0.7291	0.7292	0.7351	0.9833
0.0189	4.0	528	0.0610	0.7405	0.7360	0.7382	0.7386	0.7445	0.9846
0.0144	5.0	660	0.0702	0.7511	0.7511	0.7511	0.7511	0.7535	0.9844
0.0092	6.0	792	0.0778	0.7704	0.7366	0.7532	0.7534	0.7610	0.9843
0.0074	7.0	924	0.0825	0.7487	0.7631	0.7558	0.7563	0.7673	0.9837
0.0058	8.0	1056	0.0802	0.7603	0.7402	0.7502	0.7501	0.7515	0.9839
0.0051	9.0	1188	0.0809	0.7270	0.7541	0.7403	0.7412	0.7509	0.9834
0.0042	10.0	1320	0.0835	0.7731	0.7601	0.7665	0.7667	0.7751	0.9845
0.0033	11.0	1452	0.0847	0.7707	0.7517	0.7610	0.7612	0.7697	0.9840
0.0023	12.0	1584	0.0926	0.7559	0.7523	0.7541	0.7544	0.7605	0.9840
0.0021	13.0	1716	0.0941	0.768	0.7505	0.7591	0.7596	0.7699	0.9841
0.0018	14.0	1848	0.0980	0.7938	0.7384	0.7651	0.7653	0.7742	0.9840
0.0016	15.0	1980	0.0958	0.7755	0.7667	0.7711	0.7713	0.7770	0.9844
0.0015	16.0	2112	0.0966	0.7758	0.7637	0.7697	0.7699	0.7757	0.9845
0.0012	17.0	2244	0.0990	0.7757	0.7571	0.7663	0.7666	0.7751	0.9844
0.0012	18.0	2376	0.1001	0.7792	0.7619	0.7704	0.7707	0.7778	0.9844
0.0011	19.0	2508	0.0999	0.7801	0.7595	0.7697	0.7699	0.7766	0.9844
0.0013	19.8517	2620	0.1006	0.7800	0.7589	0.7693	0.7695	0.7763	0.9844

Framework versions

• Transformers 4.48.2
• Pytorch 2.4.1+cu121
• Datasets 3.0.2
• Tokenizers 0.21.0