En-Nso_update2

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

• Loss: 2.4199
• Bleu: 24.4776

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: 32
• eval_batch_size: 64
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 15

Training results

Training Loss	Epoch	Step	Validation Loss	Bleu
3.6661	1.0	865	3.0081	17.6871
2.7495	2.0	1730	2.7725	20.1475
2.4533	3.0	2595	2.6433	22.5433
2.3203	4.0	3460	2.5625	22.9963
2.1356	5.0	4325	2.5190	23.5696
2.0258	6.0	5190	2.4881	23.8367
1.9481	7.0	6055	2.4641	24.0611
1.8769	8.0	6920	2.4526	24.3214
1.8211	9.0	7785	2.4392	24.5300
1.7689	10.0	8650	2.4307	24.4627
1.7314	11.0	9515	2.4254	24.4936
1.7	12.0	10380	2.4243	24.4673
1.6695	13.0	11245	2.4202	24.5613
1.6562	14.0	12110	2.4200	24.4886
1.6446	15.0	12975	2.4199	24.4711

Framework versions

• Transformers 4.20.1
• Pytorch 1.12.0+cu113
• Datasets 2.3.2
• Tokenizers 0.12.1