opus-mt-en-lg-finetuned-en-to-lg

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

• Loss: 0.2360
• Bleu: 84.3078
• Gen Len: 12.3783

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: 0.0001
• train_batch_size: 32
• eval_batch_size: 32
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 10
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Bleu	Gen Len
2.3549	1.0	1253	1.5484	29.934	11.3436
1.5503	2.0	2506	1.0609	40.4943	12.2154
1.1872	3.0	3759	0.7836	48.2787	12.4935
0.9094	4.0	5012	0.5899	58.7721	12.1236
0.7069	5.0	6265	0.4598	65.7856	12.3493
0.5628	6.0	7518	0.3734	70.9367	12.5136
0.439	7.0	8771	0.3088	77.226	12.4299
0.3597	8.0	10024	0.2681	80.8605	12.4236
0.2981	9.0	11277	0.2447	83.3922	12.3797
0.2552	10.0	12530	0.2360	84.3078	12.3783

Framework versions

• Transformers 4.43.3
• Pytorch 2.3.1+cu121
• Datasets 2.20.0
• Tokenizers 0.19.1