opus-mt-en-bkm-10e6encdec

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: 1.2348
• Bleu: 11.9553
• Gen Len: 59.1988

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	1.4043	8.6981	61.258
1.5617	2.0	974	1.3571	9.509	60.8748
1.4774	3.0	1461	1.3228	10.2845	59.9297
1.4163	4.0	1948	1.2973	10.6005	58.8915
1.3617	5.0	2435	1.2759	11.0695	60.2136
1.3197	6.0	2922	1.2600	11.3475	59.4659
1.2935	7.0	3409	1.2475	11.6735	60.2644
1.263	8.0	3896	1.2408	11.8938	59.1946
1.244	9.0	4383	1.2357	12.0332	59.48
1.2308	10.0	4870	1.2348	11.9553	59.1988

Framework versions

• Transformers 4.39.2
• Pytorch 2.2.1+cu121
• Datasets 2.18.0
• Tokenizers 0.15.2