metadata

license: mit
tags:
  - generated_from_trainer
metrics:
  - bleu
  - rouge
model-index:
  - name: mbart-large-50-English_German_Translation
    results: []
language:
  - en
  - de

mbart-large-50-English_German_Translation

This model is a fine-tuned version of facebook/mbart-large-50 on the None dataset. It achieves the following results on the evaluation set:

Loss: 1.2342
Bleu: 35.5931
Rouge: {'rouge1': 0.5803386608353808, 'rouge2': 0.3939141514072567, 'rougeL': 0.5438629663406402, 'rougeLsum': 0.544153348468965}
Meteor: {'meteor': 0.5500546034636025}

Model description

Here is the link to the script I created to train this model: https://github.com/DunnBC22/NLP_Projects/blob/main/Machine%20Translation/NLP%20Translation%20Project-EN:DE.ipynb

Intended uses & limitations

This model is intended to demonstrate my ability to solve a complex problem using technology.

Training and evaluation data

Here is a the link to the page where I found this dataset: https://www.kaggle.com/datasets/hgultekin/paralel-translation-corpus-in-22-languages

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: 1

Training results

Training Loss	Epoch	Step	Validation Loss	Bleu	Rouge	Meteor
1.7738	1.0	900	1.2342	35.7436	{'rouge1': 0.5805815969432273, 'rouge2': 0.3941222478624937, 'rougeL': 0.544162316313326, 'rougeLsum': 0.5444260344836553}	{'meteor': 0.5511605039667078}

Framework versions

Transformers 4.22.2
Pytorch 1.12.1
Datasets 2.5.2
Tokenizers 0.12.1