Whisper-small-uz-V2

This model is a fine-tuned version of openai/whisper-small on the Common Voice 11.0 dataset. It achieves the following results on the evaluation set:

• Loss: 0.2628
• Wer: 23.1694

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: Use 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_steps: 500
• training_steps: 4000
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
0.4646	0.6720	1000	0.3688	32.9186
0.268	1.3441	2000	0.2925	26.4408
0.1436	2.0161	3000	0.2646	23.4813
0.1436	2.6882	4000	0.2628	23.1694

Using the Model

Use the model from the Hugging Face platform, you can use the following code:

from transformers import pipeline

# Load the model
pipe = pipeline("automatic-speech-recognition", model="tukhtashevshohruh/whisper-small-uz")

# Convert the audio file to text
audio_file = "my_audio.wav"  # Replace with the name of your own file
text = pipe(audio_file)

# Print the result
print("Text:", text['text'])

Framework versions

• Transformers 4.49.0
• Pytorch 2.5.1+cu121
• Datasets 3.3.2
• Tokenizers 0.21.0