Saltar al contenido
ModelVerificadoExterno
Descargar

Whisper Tiny Ar Quran Mix Norm

This model is a fine-tuned version of openai/whisper-tiny on the Quran dataset. It achieves the following results on the evaluation set: Loss: 0.0344 Wer Raw: 3.5416 Cer Raw: 1.0757 Wer: 3.3920 Cer: 1.0872

TipoText modelModelo baseopenai/whisper-tinyVersiónmodelPublicador@deepdmlCcalidad

Fuente del modelo

Tipo
Text model
Tamaño del archivo
144 MB
Modelo base
openai/whisper-tiny
Versión
vmodel
Actualizado
21 oct 2026
Publicador
@deepdml
Descargas
46
Fuente
Hugging Face

Descripción de la fuente

Whisper Tiny ar-quran

This model is a fine-tuned version of openai/whisper-tiny on the Quran dataset. It achieves the following results on the evaluation set:

  • Loss: 0.0344
  • Wer Raw: 3.5416
  • Cer Raw: 1.0757
  • Wer: 3.3920
  • Cer: 1.0872

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: 64
  • eval_batch_size: 64
  • seed: 42
  • optimizer: Use 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_ratio: 0.04
  • training_steps: 20000

Training results

Training LossEpochStepValidation LossWer RawCer RawWerCer
0.62090.0510000.391136.400912.340634.295211.8662
0.34870.120000.225622.42307.165121.16416.9660
0.11410.1530000.155915.59104.959114.74424.8461
0.23140.240000.130514.33735.243913.61945.1185
0.1250.2550000.099210.46223.342510.00463.3191
0.17650.360000.08488.89222.76468.46282.7475
0.30650.3570000.08018.73272.89848.27932.8643
0.0440.480000.07167.24032.19666.88572.1833
0.51880.4590000.06466.89622.17836.54492.1737
0.52670.5100000.06226.94412.24326.63652.2551
0.17950.55110000.05285.52751.76375.26631.7581
0.2220.6120000.04755.06221.62954.84461.6363
0.4870.65130000.04685.08281.70264.85521.6783
0.29440.7140000.04464.60401.43774.40041.4395
0.30010.75150000.04504.77221.51264.56091.5161
0.14930.8160000.04004.08361.24183.91111.2501
0.20010.85170000.04054.31071.42194.12531.4433
0.26170.9180000.03713.78881.14733.63211.1582
0.12780.95190000.03513.59381.10063.44671.1157
0.5541.0200000.03443.54161.07573.39201.0872

Framework versions

  • Transformers 4.48.0.dev0
  • Pytorch 2.5.1+cu121
  • Datasets 3.6.0
  • Tokenizers 0.21.0

Citation

Please cite the model using the following BibTeX entry:

@misc{deepdml/whisper-tiny-ar-quran-mix-norm,
      title={Fine-tuned Whisper tiny ASR model for speech recognition in Arabic},
      author={Jimenez, David},
      howpublished={\url{https://huggingface.co/deepdml/whisper-tiny-ar-quran-mix-norm}},
      year={2026}
    }

Fuentes

1 fuente
Hugging FacePrincipalVersión 9d8887d9ef6999f7c1f882316cce85df813b0dd9
huggingface.co/deepdml/whisper-tiny-ar-quran-mix-norm/tree/9d8887d9ef6999f7c1f882316cce85df813b0dd9

Verificado 21 oct

Artefactos del modelo

1 artefacto

model.safetensors

safetensors · 144 MB · SHA-256 5b81149ee4c5…b097 · Hugging Face

Descargar

Extractos de fuentes

3 extractos

--- library_name: transformers language: - ar license: apache-2.0 base_model: openai/whisper-tiny tags: - quran datasets: - tarteel-ai/everyayah - deepdml/tlog-clean-sf16k - tarteel-ai/EA-UD metrics: - wer model-index: - name: Whisper Tiny ar-quran results: - task: name: Automatic Speech Recognition type: automatic-speech-recognition dataset: name: Quran type: tarteel-ai/everyayah metrics: - name: Wer type: wer value: 3.3920398725229437 --- <!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. --> # Whisper Tiny ar-quran This model is a fine-tuned version of [openai/whisper-tiny](https://huggingface.co/openai/whisper-tiny) on the Quran dataset. It achieves the following results on the evaluation set: - Loss: 0.0344 - Wer Raw: 3.5416 - Cer Raw: 1.0757 - Wer: 3.3920 - Cer: 1.0872 ## 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: 64 - eval_batch_size: 64 - seed: 42 - optimizer: Use 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_ratio: 0.04 - training_steps: 20000 ### Training results | Training Loss | Epoch | Step | Validation Loss | Wer Raw | Cer Raw | Wer | Cer | |:-------------:|:-----:|:-----:|:---------------:|:-------:|:-------:|:-------:|:-------:| | 0.6209 | 0.05 | 1000 | 0.3911 | 36.4009 | 12.3406 | 34.2952 | 11.8662 | | 0.3487 | 0.1 | 2000 | 0.2256 | 22.4230 | 7.1651 | 21.1641 | 6.9660 | | 0.1141 | 0.15 | 3000 | 0.1559 | 15.5910 | 4.9591 | 14.7442 | 4.8461 | | 0.2314 | 0.2 | 4000 | 0.1305 | 14.3373 | 5.2439 | 13.6194 | 5.1185 | | 0.125 | 0.25 | 5000 | 0.0992 | 10.4622 | 3.3425 | 10.0046 | 3.3191 | | 0.1765 | 0.3 | 6000 | 0.0848 | 8.8922 | 2.7646 | 8.4628 | 2.7475 | | 0.3065 | 0.35 | 7000 | 0.0801 | 8.7327 | 2.8984 | 8.2793 | 2.8643 | | 0.044 | 0.4 | 8000 | 0.0716 | 7.2403 | 2.1966 | 6.8857 | 2.1833 | | 0.5188 | 0.45 | 9000 | 0.0646 | 6.8962 | 2.1783 | 6.5449 | 2.1737 | | 0.5267 | 0.5 | 10000 | 0.0622 | 6.9441 | 2.2432 | 6.6365 | 2.2551 | |...

Source context: 46 downloads · 1 likes · Pipeline automatic-speech-recognition · Library transformers · Repo deepdml/whisper-tiny-ar-quran-mix-norm

Source context: 675 downloads · 1 likes · Pipeline automatic-speech-recognition · Library transformers · Repo deepdml/whisper-tiny-ar-quran-mix-norm