ind-to-bbc-nmt-v3

This model is a fine-tuned version of facebook/nllb-200-distilled-600M on the nusatranslation_mt dataset. It achieves the following results on the evaluation set:

• Loss: 1.2145
• Sacrebleu: 31.0264
• Gen Len: 45.24

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: 16
• eval_batch_size: 16
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_ratio: 0.1
• num_epochs: 5
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Sacrebleu	Gen Len
4.3836	1.0	413	1.7444	26.819	45.5685
1.516	2.0	826	1.3331	30.1628	45.802
1.2184	3.0	1239	1.2480	30.7892	45.563
1.1033	4.0	1652	1.2187	31.136	45.289
1.0421	5.0	2065	1.2145	31.0264	45.24

Framework versions

• Transformers 4.41.2
• Pytorch 2.3.0+cu121
• Datasets 2.14.6
• Tokenizers 0.19.1