BART_Translation_Finetune_v0

This model is a fine-tuned version of vinai/bartpho-syllable on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.4390
• Sacrebleu: 6.4288

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: 100
• eval_batch_size: 100
• seed: 42
• optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• num_epochs: 15
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Sacrebleu
0.7274	1.0	468	0.6396	1.4512
0.628	2.0	936	0.5901	2.1680
0.5744	3.0	1404	0.5431	3.0785
0.5348	4.0	1872	0.5141	3.9172
0.5044	5.0	2340	0.4905	4.3428
0.4773	6.0	2808	0.4758	4.8562
0.4575	7.0	3276	0.4647	5.2799
0.4372	8.0	3744	0.4572	5.5700
0.429	9.0	4212	0.4499	5.8456
0.4118	10.0	4680	0.4455	6.0842
0.4056	11.0	5148	0.4424	6.2403
0.3924	12.0	5616	0.4403	6.2796
0.3858	13.0	6084	0.4396	6.3191
0.386	14.0	6552	0.4389	6.4120
0.3809	15.0	7020	0.4390	6.4288

Framework versions

• Transformers 4.48.1
• Pytorch 2.5.1+cu124
• Datasets 3.2.0
• Tokenizers 0.21.0