Skip to content
ModelListedExternal
Download

Wav2vec2 Ser

This model is a fine-tuned version of elgeish/wav2vec2-large-xlsr-53-arabic on an unknown dataset. It achieves the following results on the evaluation set: Loss: 1.1711 Accuracy: 0.5944 Macro F1: 0.5922 Uar: 0.5944

Sourcehuggingface.co/mabahboh/wav2vec2-ser/tree/0be78015388a1fc285d4d2986d9bdc637328e1c4
KindOtherBase modelelgeish/wav2vec2-large-xlsr-53-arabicVersionModelBgrade

Model source

Kind
Other
File size
602 MB
Base model
elgeish/wav2vec2-large-xlsr-53-arabic
Version
vModel
Updated
Oct 25, 2026
Source
Hugging Face

Source description

wav2vec2-ser

This model is a fine-tuned version of elgeish/wav2vec2-large-xlsr-53-arabic on an unknown dataset. It achieves the following results on the evaluation set:

  • Loss: 1.1711
  • Accuracy: 0.5944
  • Macro F1: 0.5922
  • Uar: 0.5944

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: 8
  • eval_batch_size: 8
  • 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: linear
  • lr_scheduler_warmup_steps: 0.1
  • num_epochs: 10
  • mixed_precision_training: Native AMP

Training results

Training LossEpochStepValidation LossAccuracyMacro F1Uar
2.47451.022611.30440.51460.50520.5145
2.36792.045221.22060.55820.55830.5581
2.40633.067831.19950.57280.57170.5727
2.30474.090441.19170.57470.57280.5747
2.32195.0113051.18180.58540.58430.5853
2.14586.0135661.17580.58800.58680.5880
2.15087.0158271.17310.59310.59120.5931
2.32478.0180881.17280.59240.58990.5924
2.17799.0203491.17260.59350.59090.5935
2.218310.0226101.17110.59440.59220.5944

Framework versions

  • Transformers 5.12.1
  • Pytorch 2.11.0+cu128
  • Datasets 5.0.0
  • Tokenizers 0.22.2

Sources

1 source
Hugging FacePrimaryVersion 0be78015388a1fc285d4d2986d9bdc637328e1c4
huggingface.co/mabahboh/wav2vec2-ser/tree/0be78015388a1fc285d4d2986d9bdc637328e1c4

Verified Oct 25

Model artifacts

1 artifact

model.safetensors

safetensors · 602 MB · SHA-256 b30427e6e50c…c472 · Hugging Face

Download file

Source excerpts

2 excerpts
ReadmeAug 7huggingface.co/mabahboh/wav2vec2-ser

--- library_name: transformers license: apache-2.0 base_model: elgeish/wav2vec2-large-xlsr-53-arabic tags: - generated_from_trainer metrics: - accuracy model-index: - name: wav2vec2-ser 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. --> # wav2vec2-ser This model is a fine-tuned version of [elgeish/wav2vec2-large-xlsr-53-arabic](https://huggingface.co/elgeish/wav2vec2-large-xlsr-53-arabic) on an unknown dataset. It achieves the following results on the evaluation set: - Loss: 1.1711 - Accuracy: 0.5944 - Macro F1: 0.5922 - Uar: 0.5944 ## 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: 8 - eval_batch_size: 8 - 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: linear - lr_scheduler_warmup_steps: 0.1 - num_epochs: 10 - mixed_precision_training: Native AMP ### Training results | Training Loss | Epoch | Step | Validation Loss | Accuracy | Macro F1 | Uar | |:-------------:|:-----:|:-----:|:---------------:|:--------:|:--------:|:------:| | 2.4745 | 1.0 | 2261 | 1.3044 | 0.5146 | 0.5052 | 0.5145 | | 2.3679 | 2.0 | 4522 | 1.2206 | 0.5582 | 0.5583 | 0.5581 | | 2.4063 | 3.0 | 6783 | 1.1995 | 0.5728 | 0.5717 | 0.5727 | | 2.3047 | 4.0 | 9044 | 1.1917 | 0.5747 | 0.5728 | 0.5747 | | 2.3219 | 5.0 | 11305 | 1.1818 | 0.5854 | 0.5843 | 0.5853 | | 2.1458 | 6.0 | 13566 | 1.1758 | 0.5880 | 0.5868 | 0.5880 | | 2.1508 | 7.0 | 15827 | 1.1731 | 0.5931 | 0.5912 | 0.5931 | | 2.3247 | 8.0 | 18088 | 1.1728 | 0.5924 | 0.5899 | 0.5924 | | 2.1779 | 9.0 | 20349 | 1.1726 | 0.5935 | 0.5909 | 0.5935 | | 2.2183 | 10.0 | 22610 | 1.1711 | 0.5944 | 0.5922 | 0.5944 | ### Framework versions - Transformers 5.12.1 - Pytorch 2.11.0+cu128 - Datasets 5.0.0 - Tokenizers 0.22.2

NoteAug 7huggingface.co/mabahboh/wav2vec2-ser

Source context: 4 downloads · 0 likes · Pipeline audio-classification · Library transformers · Repo mabahboh/wav2vec2-ser