scenario-non-kd-po-ner-full-mdeberta_data-univner_en66

This model is a fine-tuned version of haryoaw/scenario-TCR-NER_data-univner_en on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.1809
• Precision: 0.7243
• Recall: 0.7371
• F1: 0.7306
• Accuracy: 0.9781

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: 3e-05
• train_batch_size: 32
• eval_batch_size: 32
• seed: 66
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 30

Training results

Training Loss	Epoch	Step	Validation Loss	Precision	Recall	F1	Accuracy
0.1178	1.2755	500	0.1237	0.5196	0.5342	0.5268	0.9651
0.0471	2.5510	1000	0.1103	0.5816	0.7050	0.6373	0.9717
0.0225	3.8265	1500	0.0984	0.6793	0.7433	0.7098	0.9769
0.0131	5.1020	2000	0.1172	0.7013	0.7340	0.7172	0.9772
0.0077	6.3776	2500	0.1143	0.7055	0.7516	0.7278	0.9781
0.0057	7.6531	3000	0.1293	0.6858	0.7547	0.7186	0.9774
0.0043	8.9286	3500	0.1330	0.7348	0.7516	0.7431	0.9787
0.0031	10.2041	4000	0.1505	0.6706	0.7650	0.7147	0.9768
0.0024	11.4796	4500	0.1485	0.6998	0.7433	0.7209	0.9777
0.0019	12.7551	5000	0.1482	0.7143	0.7402	0.7270	0.9777
0.0019	14.0306	5500	0.1563	0.7140	0.7288	0.7213	0.9777
0.001	15.3061	6000	0.1596	0.7173	0.7381	0.7276	0.9777
0.0012	16.5816	6500	0.1617	0.7118	0.7236	0.7177	0.9771
0.0007	17.8571	7000	0.1688	0.7460	0.7329	0.7394	0.9785
0.0007	19.1327	7500	0.1726	0.7393	0.7340	0.7366	0.9790
0.0008	20.4082	8000	0.1685	0.7367	0.7184	0.7275	0.9784
0.0006	21.6837	8500	0.1733	0.7174	0.7464	0.7316	0.9781
0.0005	22.9592	9000	0.1719	0.7175	0.7205	0.7190	0.9775
0.0004	24.2347	9500	0.1720	0.7203	0.7464	0.7331	0.9784
0.0004	25.5102	10000	0.1748	0.7137	0.7329	0.7232	0.9780
0.0003	26.7857	10500	0.1782	0.7187	0.7433	0.7308	0.9781
0.0002	28.0612	11000	0.1827	0.7453	0.7360	0.7406	0.9790
0.0002	29.3367	11500	0.1809	0.7243	0.7371	0.7306	0.9781

Framework versions

• Transformers 4.44.2
• Pytorch 2.1.1+cu121
• Datasets 2.14.5
• Tokenizers 0.19.1