Mardiyyah/CeLLaTe-ner-2class-pubmedbert-tapt-tokenizer-adapted-spanmask-comData-lr_2

CeLLaTe-ner-2class-reinit_llrd-pubmedbert-tapt-tokenizer-adapted-spanmask-comData-lr_2.187

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: 0.0912
• Precision: 0.7753
• Recall: 0.7408
• Micro F1: 0.7577
• Weighted F1: 0.7578
• Macro F1: 0.7671
• Accuracy: 0.9841

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: 2.186715750483408e-05
• train_batch_size: 32
• eval_batch_size: 32
• seed: 2026
• 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: linear
• lr_scheduler_warmup_ratio: 0.05
• 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
0.6165	1.0	132	0.1198	0.2812	0.3927	0.3277	0.2639	0.1867	0.9564
0.0728	2.0	264	0.0566	0.7040	0.6993	0.7017	0.7019	0.7223	0.9829
0.0351	3.0	396	0.0558	0.7171	0.7120	0.7145	0.7150	0.7293	0.9827
0.0275	4.0	528	0.0584	0.7120	0.7372	0.7244	0.7253	0.7377	0.9826
0.0206	5.0	660	0.0620	0.7697	0.7156	0.7417	0.7417	0.7429	0.9833
0.0163	6.0	792	0.0679	0.7748	0.7138	0.7430	0.7433	0.7511	0.9833
0.0127	7.0	924	0.0672	0.7732	0.7174	0.7442	0.7443	0.7484	0.9835
0.0102	8.0	1056	0.0805	0.7574	0.7246	0.7406	0.7409	0.7456	0.9833
0.0081	9.0	1188	0.0758	0.7414	0.7396	0.7405	0.7407	0.7449	0.9837
0.0071	10.0	1320	0.0787	0.7596	0.7468	0.7532	0.7533	0.7577	0.9841
0.0059	11.0	1452	0.0777	0.7470	0.7456	0.7463	0.7466	0.7524	0.9836
0.0054	12.0	1584	0.0816	0.7701	0.7372	0.7533	0.7534	0.7559	0.9840
0.005	13.0	1716	0.0919	0.7881	0.7180	0.7514	0.7514	0.7620	0.9835
0.004	14.0	1848	0.0903	0.7718	0.7402	0.7557	0.7559	0.7615	0.9839
0.0037	15.0	1980	0.0908	0.7753	0.7408	0.7577	0.7578	0.7671	0.9841
0.0033	16.0	2112	0.0904	0.7605	0.7408	0.7505	0.7506	0.7555	0.9840
0.003	17.0	2244	0.0944	0.7688	0.7300	0.7489	0.7492	0.7580	0.9835
0.0026	18.0	2376	0.0946	0.7623	0.7288	0.7452	0.7455	0.7533	0.9837
0.0027	19.0	2508	0.0937	0.7658	0.7372	0.7512	0.7515	0.7587	0.9839
0.0026	19.8517	2620	0.0949	0.7727	0.7360	0.7539	0.7541	0.7612	0.9840

Framework versions

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