morten-j/fine_tuned_mBERT_old

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: 0.3681
• F1: 0.7857
• Precision: 0.8462
• Recall: 0.7333
• Accuracy: 0.8966

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: 10
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	F1	Precision	Recall	Accuracy
No log	1.0	15	0.4987	0.6875	0.6471	0.7333	0.8276
No log	2.0	30	0.4779	0.625	0.5882	0.6667	0.7931
No log	3.0	45	0.4019	0.5833	0.7778	0.4667	0.8276
No log	4.0	60	0.6146	0.6522	0.4839	1.0	0.7241
No log	5.0	75	0.3311	0.7143	0.7692	0.6667	0.8621
No log	6.0	90	0.3877	0.7568	0.6364	0.9333	0.8448
No log	7.0	105	0.3971	0.7778	0.6667	0.9333	0.8621
No log	8.0	120	0.2041	0.8966	0.9286	0.8667	0.9483
No log	9.0	135	0.2831	0.875	0.8235	0.9333	0.9310
No log	10.0	150	0.2868	0.875	0.8235	0.9333	0.9310

Framework versions

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