bangla_asr / README.md
danielbubiola's picture
update model card README.md
919dafc
|
raw
history blame
1.4 kB
metadata
tags:
  - generated_from_trainer
model-index:
  - name: bangla_asr
    results: []

bangla_asr

This model is a fine-tuned version of Harveenchadha/vakyansh-wav2vec2-bengali-bnm-200 on the None dataset. It achieves the following results on the evaluation set:

  • Loss: 1.0364
  • Wer: 0.6448

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: 0.0001
  • train_batch_size: 32
  • 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: 1000
  • num_epochs: 30
  • mixed_precision_training: Native AMP

Training results

Training Loss Epoch Step Validation Loss Wer
6.3713 15.62 500 1.0364 0.6448

Framework versions

  • Transformers 4.11.3
  • Pytorch 1.10.0+cu111
  • Datasets 1.13.3
  • Tokenizers 0.10.3