metadata

license: apache-2.0
base_model: Helsinki-NLP/opus-mt-en-ro
tags:
  - generated_from_trainer
datasets:
  - arrow
metrics:
  - bleu
model-index:
  - name: opus-mt-en-bkm-10e
    results:
      - task:
          name: Sequence-to-sequence Language Modeling
          type: text2text-generation
        dataset:
          name: arrow
          type: arrow
          config: default
          split: train
          args: default
        metrics:
          - name: Bleu
            type: bleu
            value: 6.3359

opus-mt-en-bkm-10e

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

Loss: 1.6318
Bleu: 6.3359
Gen Len: 58.0829

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: 128
eval_batch_size: 128
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	140	3.1399	0.0997	69.5849
No log	2.0	280	2.2017	1.6584	62.2953
No log	3.0	420	1.9380	3.61	59.1948
3.0959	4.0	560	1.8188	4.5827	57.0106
3.0959	5.0	700	1.7496	5.1558	56.731
3.0959	6.0	840	1.7012	5.6941	58.2515
3.0959	7.0	980	1.6683	5.964	57.0337
1.851	8.0	1120	1.6473	6.1539	57.0685
1.851	9.0	1260	1.6357	6.3233	57.5607
1.851	10.0	1400	1.6318	6.3359	58.0829

Framework versions

Transformers 4.38.2
Pytorch 2.2.1+cu121
Datasets 2.18.0
Tokenizers 0.15.2