metadata

license: apache-2.0
tags:
  - whisper
  - fine-tuned
  - malay
  - speech-to-text
datasets:
  - custom-dataset
model-index:
  - name: whisper-RMTfinetuned
    results:
      - task:
          type: automatic-speech-recognition
        dataset:
          name: Malay Audio Datasets
          type: custom
        metrics:
          - type: wer
            value: 5.6

Whisper-RMTfinetuned

This model is a fine-tuned version of OpenAI's Whisper model for Malay speech-to-text transcription.

Model Description

Base Model: OpenAI Whisper-Small
Fine-Tuned on: Malay language dataset
Intended Use: Speech recognition for Malay audio

Usage

from transformers import WhisperProcessor, WhisperForConditionalGeneration
import torch

model = WhisperForConditionalGeneration.from_pretrained("rmtariq/whisper-RMTfinetuned")
processor = WhisperProcessor.from_pretrained("rmtariq/whisper-RMTfinetuned")

audio = "/path/to/audio.wav"
input_features = processor(audio, sampling_rate=16000, return_tensors="pt").input_features

with torch.no_grad():
    predicted_ids = model.generate(input_features)

transcription = processor.batch_decode(predicted_ids, skip_special_tokens=True)[0]
print(transcription)