NIRVLab/BartEde

This model is continued pretraining from vinai/bartpho-syllable using the Ede side of NIRVLab/rhade-vietnamese-mt with a denoising/span-masking objective.

Best checkpoint

  • checkpoint: checkpoint-2676

Evaluation

  • validation loss: 0.4706411361694336
  • test loss: 0.46831414103507996
  • validation pseudo-perplexity: 1.6010203362790814
  • test pseudo-perplexity: 1.597299101073845

Notes

  • only translation["ede"] was used
  • best checkpoint selected by validation performance
Downloads last month
54
Safetensors
Model size
0.4B params
Tensor type
F32
·
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support

Model tree for NIRVLab/BartEde

Finetuned
(79)
this model

Dataset used to train NIRVLab/BartEde