COCO1033/bodo-whisper-asr-v1

Bodo Whisper ASR (v1)

This repository contains a fine-tuned Whisper Automatic Speech Recognition (ASR) model for the Bodo language (ISO 639-3: brx).

The model is trained to convert Bodo speech audio into text, supporting low-resource language research and applications.

---

🔍 Model Details

• Model name: bodo-whisper-asr-v1
• Base model: OpenAI Whisper
• Task: Automatic Speech Recognition (ASR)
• Language: Bodo (brx)
• Sampling rate: 16 kHz
• Framework: PyTorch
• Model format: SafeTensors

---

📊 Training Information

• Fine-tuned on a custom Bodo speech dataset
• Audio normalized to 16 kHz mono
• Trained using Hugging Face Transformers
• Optimized for short and medium-length utterances

(Dataset details can be added later if you decide to make them public.)

---

🚀 Usage

Using Hugging Face Transformers

from transformers import pipeline

asr = pipeline(
    "automatic-speech-recognition",
    model="COCO1033/bodo-whisper-asr-v1"
)

result = asr("audio.wav")
print(result["text"])