Frozen10-BERT-multilingual-finetuned-CEFR_ner-3000news

This model is a fine-tuned version of bert-base-multilingual-cased on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.6268
• Accuracy: 0.3626
• Precision: 0.5124
• Recall: 0.5049
• F1: 0.3803

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: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 12

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	Precision	Recall	F1
No log	1.0	132	0.8902	0.3136	0.4745	0.3572	0.2465
No log	2.0	264	0.7901	0.3298	0.4925	0.4019	0.2886
No log	3.0	396	0.7504	0.3365	0.5032	0.4185	0.2998
0.9069	4.0	528	0.7162	0.3434	0.5087	0.4450	0.3169
0.9069	5.0	660	0.6849	0.3498	0.4684	0.4536	0.3356
0.9069	6.0	792	0.6673	0.3537	0.5120	0.4640	0.3451
0.9069	7.0	924	0.6550	0.3562	0.4953	0.4866	0.3562
0.6585	8.0	1056	0.6505	0.3585	0.5117	0.4886	0.3572
0.6585	9.0	1188	0.6362	0.3601	0.5067	0.4987	0.3721
0.6585	10.0	1320	0.6317	0.3615	0.5095	0.5023	0.3773
0.6585	11.0	1452	0.6282	0.3622	0.5084	0.5052	0.3794
0.5862	12.0	1584	0.6268	0.3626	0.5124	0.5049	0.3803

Framework versions

• Transformers 4.41.1
• Pytorch 2.3.0+cu121
• Datasets 2.19.1
• Tokenizers 0.19.1