ania3000/fatjbert-morph

trainer_output

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: 0.7560
• Accuracy: 90.8827
• Sentence accuracy: 43.3962

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: 8
• eval_batch_size: 8
• seed: 42
• optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• num_epochs: 25

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	Sentence accuracy
No log	1.0	424	0.7115	83.5060	21.9340
1.328	2.0	848	0.5284	87.0493	29.2453
0.5876	3.0	1272	0.4704	88.1683	32.3113
0.3902	4.0	1696	0.4611	88.3133	33.4906
0.2749	5.0	2120	0.4642	89.3908	34.6698
0.2048	6.0	2544	0.4390	90.0539	37.5
0.2048	7.0	2968	0.4986	90.0746	37.9717
0.1499	8.0	3392	0.5226	89.9503	36.7925
0.1157	9.0	3816	0.5605	90.0332	38.4434
0.0851	10.0	4240	0.5399	90.3440	38.6792
0.0689	11.0	4664	0.6035	90.3440	42.2170
0.0499	12.0	5088	0.5832	90.5719	41.7453
0.0381	13.0	5512	0.6360	90.5097	42.2170
0.0381	14.0	5936	0.6389	90.9449	42.9245
0.0253	15.0	6360	0.6686	90.7998	42.6887
0.0198	16.0	6784	0.7092	90.5097	42.2170
0.0142	17.0	7208	0.6949	90.7791	42.6887
0.0105	18.0	7632	0.7071	90.9863	43.3962
0.0087	19.0	8056	0.7116	90.8413	43.6321
0.0087	20.0	8480	0.7549	90.7377	41.9811
0.005	21.0	8904	0.7310	90.7377	43.8679
0.0041	22.0	9328	0.7560	90.8827	43.3962

Framework versions

• Transformers 4.57.3
• Pytorch 2.11.0+cu130
• Datasets 4.0.0
• Tokenizers 0.22.2