Biobert-base-cased-v1.2-finetuned-ner-CRAFT-EN-FR-ES-IT-DE-Aug-NoEWC

This model is a fine-tuned version of StivenLancheros/Biobert-base-cased-v1.2-finetuned-ner-CRAFT-EN-FR-ES-IT-DE2-NoEWC on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.4245
• Precision: 0.7158
• Recall: 0.7763
• F1: 0.7448
• Accuracy: 0.9230

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: 16
• eval_batch_size: 16
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 20

Training results

Training Loss	Epoch	Step	Validation Loss	Precision	Recall	F1	Accuracy
0.4287	1.0	1360	0.4776	0.6125	0.7809	0.6865	0.8944
0.274	2.0	2720	0.4727	0.6370	0.7705	0.6974	0.9026
0.2219	3.0	4080	0.4405	0.6790	0.7695	0.7215	0.9134
0.1964	4.0	5440	0.4245	0.7158	0.7763	0.7448	0.9230
0.1599	5.0	6800	0.4378	0.7029	0.7732	0.7364	0.9204
0.1489	6.0	8160	0.4324	0.7018	0.7749	0.7365	0.9196
0.1342	7.0	9520	0.4534	0.7131	0.7867	0.7481	0.9216
0.12	8.0	10880	0.4532	0.7230	0.7865	0.7534	0.9243
0.1135	9.0	12240	0.4688	0.7155	0.7705	0.7420	0.9206
0.107	10.0	13600	0.4713	0.7214	0.7863	0.7524	0.9241
0.1004	11.0	14960	0.4594	0.7278	0.7823	0.7541	0.9251
0.093	12.0	16320	0.4564	0.7343	0.7937	0.7628	0.9275
0.0862	13.0	17680	0.4699	0.7242	0.7970	0.7588	0.9262
0.0788	14.0	19040	0.4745	0.7397	0.7935	0.7657	0.9285
0.0785	15.0	20400	0.4822	0.7389	0.7998	0.7682	0.9283
0.0711	16.0	21760	0.5031	0.7255	0.7964	0.7593	0.9264
0.073	17.0	23120	0.5038	0.7317	0.7976	0.7632	0.9274
0.0695	18.0	24480	0.4860	0.7411	0.8016	0.7702	0.9296
0.0649	19.0	25840	0.4978	0.7376	0.7977	0.7665	0.9285
0.0605	20.0	27200	0.4983	0.7371	0.7974	0.7661	0.9286

Framework versions

• Transformers 4.27.2
• Pytorch 1.13.0+cu117
• Datasets 2.7.1
• Tokenizers 0.13.2