xlm-large-finetuned-ner-covidmed-v5

This model is a fine-tuned version of FacebookAI/xlm-roberta-large on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.0778
• Accuracy: 0.9818
• Precision: 0.9105
• Recall: 0.9395
• F1: 0.9241
• Age Precision: 0.9692
• Age Recall: 0.9725
• Age F1-score: 0.9708
• Date Precision: 0.9832
• Date Recall: 0.9927
• Date F1-score: 0.9880
• Gender Precision: 0.9539
• Gender Recall: 0.9848
• Gender F1-score: 0.9691
• Job Precision: 0.6667
• Job Recall: 0.8208
• Job F1-score: 0.7358
• Location Precision: 0.9394
• Location Recall: 0.9532
• Location F1-score: 0.9462
• Name Precision: 0.9128
• Name Recall: 0.9214
• Name F1-score: 0.9171
• Organization Precision: 0.8692
• Organization Recall: 0.8962
• Organization F1-score: 0.8825
• Patient Id Precision: 0.9786
• Patient Id Recall: 0.9796
• Patient Id F1-score: 0.9791
• Symptom And Disease Precision: 0.8632
• Symptom And Disease Recall: 0.8944
• Symptom And Disease F1-score: 0.8785
• Transportation Precision: 0.9692
• Transportation Recall: 0.9793
• Transportation F1-score: 0.9742
• Micro avg Precision: 0.9369
• Micro avg Recall: 0.9536
• Micro avg F1-score: 0.9452
• Macro avg Precision: 0.9105
• Macro avg Recall: 0.9395
• Macro avg F1-score: 0.9241
• Weighted avg Precision: 0.9381
• Weighted avg Recall: 0.9536
• Weighted avg F1-score: 0.9457

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: 32
• eval_batch_size: 32
• seed: 42
• 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
• num_epochs: 4

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	Precision	Recall	F1	Age Precision	Age Recall	Age F1-score	Date Precision	Date Recall	Date F1-score	Gender Precision	Gender Recall	Gender F1-score	Job Precision	Job Recall	Job F1-score	Location Precision	Location Recall	Location F1-score	Name Precision	Name Recall	Name F1-score	Organization Precision	Organization Recall	Organization F1-score	Patient Id Precision	Patient Id Recall	Patient Id F1-score	Symptom And Disease Precision	Symptom And Disease Recall	Symptom And Disease F1-score	Transportation Precision	Transportation Recall	Transportation F1-score	Micro avg Precision	Micro avg Recall	Micro avg F1-score	Macro avg Precision	Macro avg Recall	Macro avg F1-score	Weighted avg Precision	Weighted avg Recall	Weighted avg F1-score
No log	1.0	158	0.0963	0.9733	0.7903	0.9007	0.8387	0.8174	0.9845	0.8932	0.9814	0.9915	0.9865	0.7897	0.9913	0.8791	0.4495	0.7457	0.5609	0.9119	0.9302	0.9210	0.8216	0.8836	0.8515	0.7558	0.8911	0.8179	0.9286	0.9791	0.9531	0.7831	0.8741	0.8261	0.6636	0.7358	0.6978	0.8717	0.9371	0.9032	0.7903	0.9007	0.8387	0.8790	0.9371	0.9059
No log	2.0	316	0.0762	0.9797	0.8970	0.9193	0.9078	0.9659	0.9725	0.9692	0.9791	0.9897	0.9844	0.9479	0.9848	0.9660	0.6392	0.7168	0.6757	0.9368	0.9514	0.9440	0.8627	0.9088	0.8851	0.8793	0.8885	0.8839	0.9752	0.9791	0.9771	0.8352	0.8477	0.8414	0.9485	0.9534	0.9509	0.9308	0.9451	0.9379	0.8970	0.9193	0.9078	0.9314	0.9451	0.9382
No log	3.0	474	0.0761	0.9812	0.9018	0.9405	0.9199	0.9468	0.9794	0.9628	0.9844	0.9933	0.9889	0.9459	0.9848	0.9650	0.6606	0.8439	0.7411	0.9222	0.9525	0.9371	0.8981	0.9151	0.9065	0.8672	0.8807	0.8739	0.9729	0.9850	0.9789	0.8555	0.8908	0.8728	0.9643	0.9793	0.9717	0.9266	0.9536	0.9399	0.9018	0.9405	0.9199	0.9279	0.9536	0.9404
0.1535	4.0	632	0.0778	0.9818	0.9105	0.9395	0.9241	0.9692	0.9725	0.9708	0.9832	0.9927	0.9880	0.9539	0.9848	0.9691	0.6667	0.8208	0.7358	0.9394	0.9532	0.9462	0.9128	0.9214	0.9171	0.8692	0.8962	0.8825	0.9786	0.9796	0.9791	0.8632	0.8944	0.8785	0.9692	0.9793	0.9742	0.9369	0.9536	0.9452	0.9105	0.9395	0.9241	0.9381	0.9536	0.9457

Framework versions

• Transformers 4.51.3
• Pytorch 2.6.0+cu124
• Datasets 3.6.0
• Tokenizers 0.21.1