j0hngou's picture
Add multilingual to the language tag (#2)
7bc5a21
metadata
language:
  - en
  - it
  - multilingual
license: apache-2.0
tags:
  - generated_from_trainer
datasets:
  - ccmatrix
model-index:
  - name: t5-small-finetuned-en-to-it
    results: []

t5-small-finetuned-en-to-it

This model is a fine-tuned version of t5-small on the ccmatrix dataset. It achieves the following results on the evaluation set:

  • Loss: 2.0188

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: 16
  • eval_batch_size: 16
  • seed: 42
  • optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • lr_scheduler_type: linear
  • num_epochs: 15
  • mixed_precision_training: Native AMP

Training results

Training Loss Epoch Step Validation Loss
2.5524 1.0 750 2.2315
2.4839 2.0 1500 2.1932
2.4654 3.0 2250 2.1637
2.4001 4.0 3000 2.1352
2.3966 5.0 3750 2.1122
2.3537 6.0 4500 2.0921
2.3427 7.0 5250 2.0746
2.316 8.0 6000 2.0614
2.301 9.0 6750 2.0488
2.2813 10.0 7500 2.0403
2.2691 11.0 8250 2.0325
2.2561 12.0 9000 2.0265
2.258 13.0 9750 2.0217
2.2447 14.0 10500 2.0199
2.2432 15.0 11250 2.0188

Framework versions

  • Transformers 4.22.1
  • Pytorch 1.12.1+cu113
  • Datasets 2.5.1
  • Tokenizers 0.12.1