ko_en_nllb-200-distilled-1.3B

This model is a fine-tuned version of facebook/nllb-200-distilled-1.3B on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.2977
• Bleu: 0.4125
• Gen Len: 26.559

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: 64
• eval_batch_size: 64
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 256
• 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
• lr_scheduler_warmup_steps: 500
• num_epochs: 10

Training results

Training Loss	Epoch	Step	Validation Loss	Bleu	Gen Len
0.4558	1.1212	500	0.4154	0.3556	26.3133
0.272	2.2424	1000	0.3005	0.3987	26.5558
0.2464	3.3636	1500	0.2925	0.4032	26.4593
0.2257	4.4848	2000	0.2905	0.4103	26.5054
0.2192	5.6061	2500	0.2916	0.4104	26.595
0.2021	6.7273	3000	0.2935	0.411	26.5823
0.1931	7.8485	3500	0.2959	0.4122	26.6143
0.1862	8.9697	4000	0.2977	0.4125	26.559

Framework versions

• Transformers 4.47.0
• Pytorch 2.5.1+cu124
• Datasets 3.1.0
• Tokenizers 0.21.0