Whisper Small Vi - Nhi Le

This model is a fine-tuned version of openai/whisper-small on the Vietnamese ASR Custom Corpus dataset. It achieves the following results on the evaluation set:

  • Loss: 0.9983
  • Wer: 245.2381

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: 1e-05
  • train_batch_size: 16
  • eval_batch_size: 8
  • seed: 42
  • optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • lr_scheduler_type: linear
  • lr_scheduler_warmup_steps: 2
  • training_steps: 10

Training results

Training Loss Epoch Step Validation Loss Wer
2.099 1.0 2 1.9868 26.1905
1.779 2.0 4 1.5173 257.1429
1.1783 3.0 6 1.2374 245.2381
0.964 4.0 8 1.0648 245.2381
0.89 5.0 10 0.9983 245.2381

Framework versions

  • Transformers 4.36.0
  • Pytorch 2.1.0+cpu
  • Datasets 2.15.0
  • Tokenizers 0.15.0
Downloads last month
-
Safetensors
Model size
0.2B params
Tensor type
F32
·
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support

Model tree for nhihlle/whisper-small-vi

Finetuned
(3287)
this model