Beta
Explore
Marketplace
Neural Labs
Chat
Wallet
Docs
whisper-dspData-v5 – AI Model by anushaporwal | AlphaNeural AI
You can deploy this model and start earning money today!
anushaporwal
/
whisper-dspData-v5
like
0
transformers
tensorboard
safetensors
whisper
automatic-speech-recognition
generated_from_trainer
openai/whisper-medium
finetune
apache-2.0
endpoints_compatible
us
Views
No views yet
Model card
Files and Versions
Community
API
Deploy
whisper-dspData-v5
This model is a fine-tuned version of
openai/whisper-medium
on an unknown dataset. It achieves the following results on the evaluation set:
Loss: 0.1537
Wer: 6.2767
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
training_steps: 1000
mixed_precision_training: Native AMP
Training results
Training Loss
Epoch
Step
Validation Loss
Wer
0.0033
2.1605
350
0.1496
6.4117
0.0002
4.3210
700
0.1537
6.2767
Framework versions
Transformers 4.44.2
Pytorch 2.4.0
Datasets 2.21.0
Tokenizers 0.19.1
