CeLLaTe-ner-3class-bioformer16l-baseline

This model is a fine-tuned version of bioformers/bioformer-16l on the OTAR3088/CeLLaTe-ner-3class-iob_final dataset. It achieves the following results on the evaluation set:

• Loss: 0.1046
• Precision: 0.7689
• Recall: 0.7562
• Micro F1: 0.7625
• Weighted F1: 0.7625
• Macro F1: 0.7538
• Accuracy: 0.9801

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: 3407
• gradient_accumulation_steps: 2
• total_train_batch_size: 32
• 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: 0.01
• 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
0.4159	1.0	263	0.1063	0.3976	0.4541	0.4240	0.4262	0.3937	0.9635
0.08	2.0	526	0.0766	0.7320	0.6550	0.6913	0.6899	0.6770	0.9776
0.0486	3.0	789	0.0722	0.7410	0.6940	0.7167	0.7160	0.7085	0.9788
0.0354	4.0	1052	0.0810	0.7692	0.6798	0.7217	0.7172	0.7018	0.9786
0.0271	5.0	1315	0.0821	0.7597	0.7034	0.7305	0.7300	0.7210	0.9791
0.0216	6.0	1578	0.0909	0.7005	0.6908	0.6956	0.6966	0.6900	0.9769
0.0186	7.0	1841	0.0895	0.7309	0.7188	0.7248	0.7255	0.7172	0.9779
0.0147	8.0	2104	0.1000	0.7386	0.7022	0.7200	0.7199	0.7126	0.9783
0.0125	9.0	2367	0.0911	0.7454	0.7519	0.7486	0.7487	0.7389	0.9787
0.0115	10.0	2630	0.0952	0.7400	0.7184	0.7290	0.7292	0.7196	0.9788
0.0097	11.0	2893	0.0996	0.7411	0.7271	0.7340	0.7337	0.7250	0.9787
0.0085	12.0	3156	0.1020	0.7720	0.7361	0.7536	0.7536	0.7456	0.9802
0.0079	13.0	3419	0.1087	0.7121	0.7200	0.7160	0.7167	0.7092	0.9776
0.0075	14.0	3682	0.1087	0.7590	0.7294	0.7439	0.7442	0.7336	0.9793
0.0065	15.0	3945	0.1047	0.7689	0.7562	0.7625	0.7625	0.7538	0.9801
0.0058	16.0	4208	0.1107	0.7373	0.7373	0.7373	0.7382	0.7314	0.9787
0.0053	17.0	4471	0.1142	0.7494	0.7361	0.7427	0.7433	0.7355	0.9792
0.005	18.0	4734	0.1154	0.7324	0.7361	0.7342	0.7352	0.7273	0.9785
0.0047	19.0	4997	0.1158	0.7405	0.7397	0.7401	0.7407	0.7325	0.9790
0.0048	20.0	5260	0.1152	0.7524	0.7444	0.7484	0.7487	0.7397	0.9794

Framework versions

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