dimasichsanul/best-mbert-reranker

results

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: 0.4003
• Accuracy: 0.8589
• F1: 0.7308
• Precision: 0.7238
• Recall: 0.7379

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: 16
• eval_batch_size: 16
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 100
• num_epochs: 5

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	F1	Precision	Recall
No log	0.2623	16	0.6597	0.7406	0.0	0.0	0.0
No log	0.5246	32	0.5547	0.7406	0.0	0.0	0.0
No log	0.7869	48	0.5144	0.7406	0.0	0.0	0.0
No log	1.0492	64	0.4658	0.8237	0.5205	0.8837	0.3689
No log	1.3115	80	0.4164	0.8338	0.7	0.6581	0.7476
No log	1.5738	96	0.3812	0.8212	0.6872	0.6290	0.7573
No log	1.8361	112	0.3799	0.8564	0.6705	0.8286	0.5631
No log	2.0984	128	0.3736	0.8111	0.6725	0.6111	0.7476
No log	2.3607	144	0.3726	0.8564	0.7047	0.7556	0.6602
No log	2.6230	160	0.4651	0.7456	0.6456	0.5055	0.8932
No log	2.8852	176	0.3592	0.8413	0.7070	0.6786	0.7379
No log	3.1475	192	0.3633	0.8514	0.7035	0.7292	0.6796
No log	3.4098	208	0.4381	0.8086	0.6984	0.5906	0.8544
No log	3.6721	224	0.4114	0.8338	0.7080	0.6504	0.7767
No log	3.9344	240	0.4588	0.8186	0.7025	0.6115	0.8252
No log	4.1967	256	0.3795	0.8615	0.7291	0.74	0.7184
No log	4.4590	272	0.4418	0.8262	0.7113	0.625	0.8252
No log	4.7213	288	0.3962	0.8489	0.7170	0.6972	0.7379
No log	4.9836	304	0.4003	0.8589	0.7308	0.7238	0.7379

Framework versions

• Transformers 4.41.2
• Pytorch 2.3.0+cu121
• Tokenizers 0.19.1