Beta
Explore
Marketplace
Neural Labs
Playground
Wallet
Docs
whisper-base-arabic-v2.1 – AI Model by shredder-31 | AlphaNeural AI
You can deploy this model and start earning money today!
shredder-31
/
whisper-base-arabic-v2.1
like
0
transformers
tensorboard
safetensors
whisper
automatic-speech-recognition
generated_from_trainer
openai/whisper-base
finetune
apache-2.0
endpoints_compatible
us
Views
No views yet
Model card
Files and Versions
Community
API
Deploy
whisper-base-arabic-base-v2.1
This model is a fine-tuned version of
openai/whisper-base
on an unknown dataset. It achieves the following results on the evaluation set:
Loss: 0.5048
Wer: 0.5558
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: 32
eval_batch_size: 32
seed: 42
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_ratio: 0.01
num_epochs: 4
mixed_precision_training: Native AMP
Training results
Training Loss
Epoch
Step
Validation Loss
Wer
0.7955
1.0
1476
0.6025
0.6847
0.6122
2.0
2952
0.5387
0.6075
0.5065
3.0
4428
0.5126
0.5567
0.4651
4.0
5904
0.5048
0.5558
Framework versions
Transformers 4.53.3
Pytorch 2.6.0+cu124
Datasets 3.6.0
Tokenizers 0.21.2
