Af-En_update_pc

This model is a fine-tuned version of Helsinki-NLP/opus-mt-af-en on the None dataset. It achieves the following results on the evaluation set:

• Loss: 1.8319
• Model Preparation Time: 0.0
• Bleu: 50.6655

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: 8
• eval_batch_size: 16
• seed: 42
• optimizer: Use OptimizerNames.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: 15
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Model Preparation Time	Bleu
1.5372	1.0	5105	1.9631	0.0	36.5658
1.2529	2.0	10210	1.8745	0.0	46.3922
1.0324	3.0	15315	1.7872	0.0	47.6135
0.8459	4.0	20420	1.6956	0.0	49.7908
0.6545	5.0	25525	1.6746	0.0	50.3903
0.5743	6.0	30630	1.6951	0.0	50.8569
0.4897	7.0	35735	1.6863	0.0	50.6211
0.4328	8.0	40840	1.7113	0.0	50.6402
0.3955	9.0	45945	1.7482	0.0	50.7567
0.3565	10.0	51050	1.7655	0.0	50.8257
0.3275	11.0	56155	1.7988	0.0	50.5494
0.2696	12.0	61260	1.8035	0.0	50.6981
0.2635	13.0	66365	1.8194	0.0	50.6104
0.2526	14.0	71470	1.8190	0.0	50.6222
0.2338	15.0	76575	1.8319	0.0	50.6658

Framework versions

• Transformers 4.51.3
• Pytorch 2.6.0
• Datasets 3.5.0
• Tokenizers 0.21.1