Ir para o conteúdo
ModelListadosExterno
Baixar

Whisper Large v3 Turbo Persian Farsi Unified Cleaned

This model is a fine-tuned version of openai/whisper-large-v3-turbo on the None dataset. It achieves the following results on the evaluation set: Loss: 0.1349 Wer: 28.5195

Fontehuggingface.co/tivanai2020/whisper-large-v3-turbo-persian-farsi-unified-cleaned/tree/3b788b6f89b3ee811eadb38512dcf05be4670167
TipoLoRAModelo baseopenai/whisper-large-v3-turboVersãoadapter_modelBnota

Fonte do modelo

Tipo
LoRA
Tamanho do arquivo
50,0 MB
Modelo base
openai/whisper-large-v3-turbo
Versão
vadapter_model
Atualizado
27 de out. de 2026
Downloads
6
Fonte
Hugging Face

Descrição da fonte

whisper-large-v3-turbo-persian-farsi-unified-cleaned

This model is a fine-tuned version of openai/whisper-large-v3-turbo on the None dataset. It achieves the following results on the evaluation set:

  • Loss: 0.1349
  • Wer: 28.5195

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: 8
  • eval_batch_size: 4
  • seed: 42
  • gradient_accumulation_steps: 2
  • total_train_batch_size: 16
  • optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
  • lr_scheduler_type: cosine
  • lr_scheduler_warmup_steps: 50
  • training_steps: 9561

Training results

Training LossEpochStepValidation LossWer
0.48550.15695000.250538.5710
0.43460.313710000.220836.7999
0.45280.470615000.200735.6797
0.38200.627520000.189434.6957
0.37470.784325000.178633.3182
0.33310.941230000.173333.6966
0.27801.097935000.165832.0315
0.27581.254740000.160031.1384
0.27511.411645000.153930.3512
0.22851.568550000.150930.3512
0.23501.725355000.145929.9728
0.25691.882260000.142329.4127
0.17832.038965000.140929.3975
0.17792.195870000.139628.9585
0.19072.352675000.136628.9585
0.16982.509580000.136628.5801
0.21042.666485000.135228.6709
0.15242.823290000.134828.4741
0.17332.980195000.134928.4741
0.16262.999295610.134928.5195

Framework versions

  • PEFT 0.19.1
  • Transformers 5.13.1
  • Pytorch 2.11.0+cu128
  • Datasets 4.0.0
  • Tokenizers 0.22.2

Fontes

1 fonte
Hugging FacePrimárioVersão 3b788b6f89b3ee811eadb38512dcf05be4670167
huggingface.co/tivanai2020/whisper-large-v3-turbo-persian-farsi-unified-cleaned/tree/3b788b6f89b3ee811eadb38512dcf05be4670167

Verificado 27 de out.

Artefatos de modelo

1 artefato

adapter_model.safetensors

safetensors · 50,0 MB · SHA-256 36bf99dbb1f3…149a · Hugging Face

Download file

Trechos de fonte

3 trechos
Readme8 de ago.huggingface.co/tivanai2020/whisper-large-v3-turbo-persian-farsi-unified-cleaned

--- library_name: peft license: mit base_model: openai/whisper-large-v3-turbo tags: - base_model:adapter:openai/whisper-large-v3-turbo - lora - transformers metrics: - wer model-index: - name: whisper-large-v3-turbo-persian-farsi-unified-cleaned results: [] --- <!-- 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-large-v3-turbo-persian-farsi-unified-cleaned This model is a fine-tuned version of [openai/whisper-large-v3-turbo](https://huggingface.co/openai/whisper-large-v3-turbo) on the None dataset. It achieves the following results on the evaluation set: - Loss: 0.1349 - Wer: 28.5195 ## 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: 8 - eval_batch_size: 4 - seed: 42 - gradient_accumulation_steps: 2 - total_train_batch_size: 16 - optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments - lr_scheduler_type: cosine - lr_scheduler_warmup_steps: 50 - training_steps: 9561 ### Training results | Training Loss | Epoch | Step | Validation Loss | Wer | |:-------------:|:------:|:----:|:---------------:|:-------:| | 0.4855 | 0.1569 | 500 | 0.2505 | 38.5710 | | 0.4346 | 0.3137 | 1000 | 0.2208 | 36.7999 | | 0.4528 | 0.4706 | 1500 | 0.2007 | 35.6797 | | 0.3820 | 0.6275 | 2000 | 0.1894 | 34.6957 | | 0.3747 | 0.7843 | 2500 | 0.1786 | 33.3182 | | 0.3331 | 0.9412 | 3000 | 0.1733 | 33.6966 | | 0.2780 | 1.0979 | 3500 | 0.1658 | 32.0315 | | 0.2758 | 1.2547 | 4000 | 0.1600 | 31.1384 | | 0.2751 | 1.4116 | 4500 | 0.1539 | 30.3512 | | 0.2285 | 1.5685 | 5000 | 0.1509 | 30.3512 | | 0.2350 | 1.7253 | 5500 | 0.1459 | 29.9728 | | 0.2569 | 1.8822 | 6000 | 0.1423 | 29.4127 | | 0.1783 | 2.0389 | 6500 | 0.1409 | 29.3975 | | 0.1779 | 2.1958 | 7000 | 0.1396 | 28.9585 | | 0.1907 | 2.3526 | 7500 | 0.1366 | 28.9585 | | 0.1698 | 2.5095 | 8000 | 0.1366 | 28.5801 | | 0.2104 | 2.6664 | 8500 | 0.1352 | 28.6709 | | 0.1524 | 2.8232 | 9000 | 0.1348 | 28.4741 | | 0.1733 | 2.9801 | 9500 | 0.1349 | 28.4741 | |...

Note27 de out.huggingface.co/tivanai2020/whisper-large-v3-turbo-persian-farsi-unified-cleaned

Source context: 6 downloads · 0 likes · Library peft · Repo tivanai2020/whisper-large-v3-turbo-persian-farsi-unified-cleaned

Note8 de ago.huggingface.co/tivanai2020/whisper-large-v3-turbo-persian-farsi-unified-cleaned

Source context: 151 downloads · 0 likes · Library peft · Repo tivanai2020/whisper-large-v3-turbo-persian-farsi-unified-cleaned