mehdie/fine_tuned_mBERT

fine_tuned_bert

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

• Loss: 0.1259
• F1: 0.8182
• F5: 0.8326
• Precision: 0.7826
• Recall: 0.8571

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: 5e-05
• train_batch_size: 32
• eval_batch_size: 32
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 6
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	F1	F5	Precision	Recall
No log	1.0	65	0.2964	0.0	0.0	0.0	0.0
No log	2.0	130	0.2682	0.4737	0.4081	0.8182	0.3333
No log	3.0	195	0.2208	0.65	0.7421	0.4906	0.9630
No log	4.0	260	0.1924	0.7273	0.7816	0.6154	0.8889
No log	5.0	325	0.1246	0.8727	0.8788	0.8571	0.8889
No log	6.0	390	0.1142	0.8519	0.8519	0.8519	0.8519

Framework versions

• Transformers 4.38.1
• Pytorch 2.2.1+cu121
• Datasets 2.17.0
• Tokenizers 0.15.2