Muneeba23/whisper-small-en

Storymation-whisper Fine-Tuned Model

This model is a fine-tuned version of openai/whisper-small on the Common Voice 11.0 dataset.

Model Usage

!pip install transformers accelerate gradio
from transformers import pipeline
import gradio as gr

# Load the Whisper model
model = "Muneeba23/whisper-small-en"
pipe = pipeline(model=model)

# Define the transcribe function
def transcribe(audio):
    text = pipe(audio)["text"]
    return text

# Create the Gradio interface
iface = gr.Interface(
    fn=transcribe,
    inputs=gr.Audio(type="filepath"), 
    outputs="text",
    title="Whisper Small",
    description="Real-time Demo. Hurrah!!"
)

# Launch the interface
iface.launch()

Intended uses & limitations

For a average audio prompt of 5 secs the latency observed was 40 secs.

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: 4
• eval_batch_size: 2
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 16
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 500
• training_steps: 3

Training results

• global_step=3,
• training_loss=5.196450551350911,
• WER = 30% for 8 hours of training

Framework versions

• Transformers 4.45.2
• Pytorch 2.4.1+cu121
• Datasets 3.0.1
• Tokenizers 0.20.0