bert-base-multilingual-uncased-classification

This model is a fine-tuned version of google-bert/bert-base-multilingual-uncased on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.2025
• Accuracy: 0.9563

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: 25

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
No log	1.0	95	0.2181	0.9523
No log	2.0	190	0.2287	0.9543
No log	3.0	285	0.2135	0.9523
No log	4.0	380	0.1975	0.9523
No log	5.0	475	0.2015	0.9563
0.0475	6.0	570	0.1930	0.9523
0.0475	7.0	665	0.1784	0.9563
0.0475	8.0	760	0.2306	0.9563
0.0475	9.0	855	0.2177	0.9543
0.0475	10.0	950	0.1579	0.9563
0.0353	11.0	1045	0.1645	0.9563
0.0353	12.0	1140	0.2233	0.9563
0.0353	13.0	1235	0.1985	0.9523
0.0353	14.0	1330	0.1932	0.9543
0.0353	15.0	1425	0.2461	0.9563
0.0312	16.0	1520	0.1834	0.9563
0.0312	17.0	1615	0.1821	0.9543
0.0312	18.0	1710	0.1985	0.9563
0.0312	19.0	1805	0.1984	0.9583
0.0312	20.0	1900	0.2036	0.9583
0.0312	21.0	1995	0.1957	0.9563
0.0264	22.0	2090	0.1996	0.9563
0.0264	23.0	2185	0.2041	0.9583
0.0264	24.0	2280	0.2022	0.9583
0.0264	25.0	2375	0.2025	0.9563

Framework versions

• Transformers 4.37.2
• Pytorch 2.1.0+cu121
• Datasets 2.17.1
• Tokenizers 0.15.2