en-es

This model is a fine-tuned version of facebook/mbart-large-50 on the CarlaLlavador/Ted dataset. It achieves the following results on the evaluation set:

• Loss: 1.4216
• Bleu: 21.0129
• Gen Len: 27.4946

Model description

Este modelo ha sido entrenado a partir del dataset CarlaLlavador/Ted, configurado a partir de un corpus paralelo alineado frase a frase formado por las transcripciones en inglés y español de tres charlas Ted.

Intended uses & limitations

Es un modelo entrenado con un dataset muy pequeño, lo cual limita las posibilidades del modelo, por lo que se debe tener en cuenta.

Training and evaluation data

More information needed

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

• learning_rate: 5.6e-05
• train_batch_size: 8
• eval_batch_size: 8
• seed: 42
• optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• num_epochs: 2

Training results

Training Loss	Epoch	Step	Validation Loss	Bleu	Gen Len
No log	1.0	93	1.9252	16.4219	19.8859
No log	2.0	186	1.4216	21.0129	27.4946

Framework versions

• Transformers 4.49.0
• Pytorch 2.6.0+cu124
• Datasets 3.4.1
• Tokenizers 0.21.1