Hugging Face's logo

numan98
/
whisper-tiny-nextayah-v3

Automatic Speech Recognition
Transformers
TensorBoard
Safetensors
Arabic
whisper
Generated from Trainer
Model card Files Metrics Community
whisper-tiny-nextayah-v3 / README.md
numan98's picture
numan98
End of training
58bc570 verified
preview code
|
raw
history blame contribute delete
1.65 kB
metadata
library_name: transformers
language:
  - ar
license: apache-2.0
base_model: tarteel-ai/whisper-tiny-ar-quran
tags:
  - generated_from_trainer
datasets:
  - numan98/synth-incorrect-verses
model-index:
  - name: Nextayah Tiny Whisper Finetuned
    results: []

Nextayah Tiny Whisper Finetuned

This model is a fine-tuned version of tarteel-ai/whisper-tiny-ar-quran on the Synthetic Incorrect Verses dataset. It achieves the following results on the evaluation set:

  • eval_loss: 0.1059
  • eval_wer: 16.4345
  • eval_runtime: 49.0992
  • eval_samples_per_second: 2.037
  • eval_steps_per_second: 0.265
  • epoch: 68.9825
  • step: 2000

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: 16
  • eval_batch_size: 8
  • seed: 42
  • gradient_accumulation_steps: 2
  • total_train_batch_size: 32
  • 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
  • lr_scheduler_warmup_steps: 750
  • training_steps: 2500
  • mixed_precision_training: Native AMP

Framework versions

  • Transformers 4.48.0
  • Pytorch 2.5.1+cu121
  • Datasets 3.2.0
  • Tokenizers 0.21.0