nllb-dry-run

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

• Loss: 3.6877
• Bleu: 4.2221
• Chrf++: 17.8408

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: 8e-05
• train_batch_size: 2
• eval_batch_size: 4
• 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
• training_steps: 200

Training results

Training Loss	Epoch	Step	Validation Loss	Bleu	Chrf++
3.9222	0.3125	10	3.4697	2.3857	14.2057
3.7650	0.625	20	3.3382	3.7821	18.4207
3.8782	0.9375	30	3.3116	5.8430	19.1154
2.3540	1.25	40	3.2531	3.6516	19.0412
2.5031	1.5625	50	3.5119	3.4640	18.3757
2.3598	1.875	60	3.3649	3.8286	19.9150
2.1185	2.1875	70	3.6434	2.4977	18.1167
1.5931	2.5	80	3.4119	2.8267	16.8880
1.3210	2.8125	90	3.6877	4.2221	17.8408

Framework versions

• Transformers 5.5.4
• Pytorch 2.8.0+cu128
• Datasets 4.8.4
• Tokenizers 0.22.2