Beta
Explore
Marketplace
Neural Labs
Chat
Wallet
Docs
whisper-base-Withaq-V1 – AI Model by naiftamia | AlphaNeural AI
You can deploy this model and start earning money today!
naiftamia
/
whisper-base-Withaq-V1
like
0
transformers
tensorboard
safetensors
whisper
automatic-speech-recognition
hf-asr-leaderboard
generated_from_trainer
hi
SADA
openai/whisper-base
finetune
apache-2.0
endpoints_compatible
us
Views
No views yet
Model card
Files and Versions
Community
API
Deploy
whisper-base-Withaq-V1
This model is a fine-tuned version of
openai/whisper-base
on the SADA - Arabic dataset.
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
Framework versions
Transformers 4.39.0.dev0
Pytorch 2.1.0+cu121
Datasets 2.17.1
Tokenizers 0.15.2
