ru-kbd_lat-t5-small / README.md
anzorq's picture
Update README.md
7f9d4ce
|
raw
history blame
1.81 kB
metadata
language:
  - ru
  - kbd
license: mit
tags:
  - generated_from_trainer
datasets:
  - anzorq/kbd_lat-ru
metrics:
  - bleu
model-index:
  - name: tst-translation
    results:
      - task:
          name: Translation
          type: translation
        dataset:
          name: anzorq/kbd_lat-ru anzorq--kbd-ru
          type: anzorq/kbd_lat-ru
          args: anzorq--kbd-ru
        metrics:
          - name: Bleu
            type: bleu
            value: 12.649
widget:
  - text: 'ru->kbd: Я иду домой.'
    example_title: Я иду домой.
  - text: 'ru->kbd: Дети играют во дворе.'
    example_title: Дети играют во дворе.
  - text: 'ru->kbd: Сколько тебе лет?'
    example_title: Сколько тебе лет?

tst-translation

This model is a fine-tuned version of anzorq/kbd_lat-835k_ru-3M_t5-small on the anzorq/kbd_lat-ru anzorq--kbd-ru dataset. It achieves the following results on the evaluation set:

  • Loss: 2.6000
  • Bleu: 12.649
  • Gen Len: 11.8018

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: 5e-05
  • train_batch_size: 30
  • eval_batch_size: 30
  • seed: 42
  • optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • lr_scheduler_type: linear
  • num_epochs: 35.0

Training results

Framework versions

  • Transformers 4.21.0.dev0
  • Pytorch 1.10.0+cu113
  • Datasets 1.16.0
  • Tokenizers 0.12.1