Hugging Face's logo

golesheed
/
whisper-large-v2-fa

Automatic Speech Recognition
Transformers
TensorBoard
Safetensors
Persian
whisper
Generated from Trainer
Inference Endpoints
Model card Files Metrics Community
whisper-large-v2-fa / README.md
golesheed's picture
golesheed
End of training
08b4cdc verified
preview code
|
raw
history blame contribute delete
1.64 kB
metadata
language:
  - fa
license: apache-2.0
base_model: openai/whisper-large-v2
tags:
  - generated_from_trainer
metrics:
  - wer
model-index:
  - name: Whisper Large Fa
    results: []

Whisper Large Fa

This model is a fine-tuned version of openai/whisper-large-v2 on the None dataset. It achieves the following results on the evaluation set:

  • Loss: 0.2511
  • Wer: 52.3497

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: 500
  • training_steps: 4000
  • mixed_precision_training: Native AMP

Training results

Training Loss Epoch Step Validation Loss Wer
0.1999 0.43 1000 0.3631 55.5243
0.1391 0.86 2000 0.2965 47.4574
0.0719 1.29 3000 0.2725 54.5863
0.0611 1.72 4000 0.2511 52.3497

Framework versions

  • Transformers 4.38.0.dev0
  • Pytorch 2.1.0+cu121
  • Datasets 2.14.6
  • Tokenizers 0.15.0