lab1_finetuning

This model is a fine-tuned version of Helsinki-NLP/opus-mt-en-fr on the kde4 dataset. It achieves the following results on the evaluation set:

• Loss: 1.0255
• Model Preparation Time: 0.0056
• Bleu: 48.8948

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: 16
• eval_batch_size: 32
• 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: 5000
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Model Preparation Time	Bleu
1.4007	0.0476	500	1.2424	0.0056	45.5942
1.1468	0.0952	1000	1.1651	0.0056	46.9449
1.0415	0.1427	1500	1.1203	0.0056	47.6958
1.1744	0.1903	2000	1.0877	0.0056	44.0503
1.1876	0.2379	2500	1.0665	0.0056	48.6443
1.1702	0.2855	3000	1.0510	0.0056	47.1173
1.0369	0.3330	3500	1.0385	0.0056	48.8846
1.1668	0.3806	4000	1.0325	0.0056	49.0365
1.1351	0.4282	4500	1.0279	0.0056	48.8962
1.0436	0.4758	5000	1.0255	0.0056	49.0433

Framework versions

• Transformers 4.57.6
• Pytorch 2.10.0+cu128
• Datasets 3.6.0
• Tokenizers 0.22.2