opus-mt-en-bkm-10e10encdec

This model is a fine-tuned version of kalese/opus-mt-en-bkm-10e64 on the arrow dataset. It achieves the following results on the evaluation set:

• Loss: 1.2361
• Bleu: 12.0707
• Gen Len: 59.4315

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

Training results

Training Loss	Epoch	Step	Validation Loss	Bleu	Gen Len
No log	1.0	487	1.3987	8.8903	60.2246
1.5613	2.0	974	1.3545	9.5185	60.2484
1.4784	3.0	1461	1.3220	10.1143	59.757
1.4123	4.0	1948	1.2958	10.6244	59.5823
1.3625	5.0	2435	1.2744	11.0208	59.4767
1.3202	6.0	2922	1.2603	11.3956	59.1844
1.2875	7.0	3409	1.2498	11.6882	60.2625
1.2628	8.0	3896	1.2414	11.8521	59.6434
1.2431	9.0	4383	1.2381	11.9893	59.6424
1.2245	10.0	4870	1.2361	12.0707	59.4315

Framework versions

• Transformers 4.39.1
• Pytorch 2.2.1+cu121
• Datasets 2.18.0
• Tokenizers 0.15.2