thainq107/en-vi-mbart50

en-vi-mabrt50

This model is a fine-tuned version of facebook/mbart-large-50-many-to-many-mmt on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 1.2825
• Bleu: 34.6635

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: Use 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: 3

Training results

Training Loss	Epoch	Step	Bleu	Validation Loss
1.228	0.2400	1000	33.4010	1.3419
1.2022	0.4800	2000	34.0752	1.3063
1.1771	0.7199	3000	34.1612	1.2806
1.1607	0.9599	4000	34.3856	1.2582
0.9698	1.1999	5000	34.3075	1.2860
0.9298	1.4399	6000	34.4419	1.2671
0.9282	1.6799	7000	34.7962	1.2552
0.9174	1.9198	8000	34.7904	1.2516
0.8538	2.1598	9000	1.3000	34.3462
0.822	2.3998	10000	1.2953	34.3928
0.8206	2.6398	11000	1.2834	34.4372
0.8177	2.8798	12000	1.2825	34.6635

Framework versions

• Transformers 4.49.0
• Pytorch 2.6.0+cu124
• Datasets 2.20.0
• Tokenizers 0.21.0