README.md

metadata

tags:
  - generated_from_trainer
datasets:
  - mt_eng_vietnamese
metrics:
  - bleu
model-index:
  - name: t5vi-finetuned-en-to-vi
    results:
      - task:
          name: Sequence-to-sequence Language Modeling
          type: text2text-generation
        dataset:
          name: mt_eng_vietnamese
          type: mt_eng_vietnamese
          args: iwslt2015-en-vi
        metrics:
          - name: Bleu
            type: bleu
            value: 13.547

t5vi-finetuned-en-to-vi

This model is a fine-tuned version of imthanhlv/t5vi on the mt_eng_vietnamese dataset. It achieves the following results on the evaluation set:

• Loss: 1.3827
• Bleu: 13.547
• Gen Len: 17.3719

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: 20
• eval_batch_size: 20
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 5
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Bleu	Gen Len
1.8026	1.0	6666	1.5907	10.9756	17.3231
1.6217	2.0	13332	1.4635	12.375	17.3444
1.5087	3.0	19998	1.4131	13.1828	17.3924
1.4446	4.0	26664	1.3915	13.5217	17.3617
1.4076	5.0	33330	1.3827	13.547	17.3719

Framework versions

• Transformers 4.19.1
• Pytorch 1.11.0+cu113
• Datasets 2.2.1
• Tokenizers 0.12.1