northtechai/northstt-stageb-300

This is a fine-tuned version of openai/whisper-large-v3 using LoRA (Low-Rank Adaptation) for Turkish speech recognition.

Model Details

• Base Model: openai/whisper-large-v3
• Fine-tuning Method: LoRA (Low-Rank Adaptation)
• Language: Turkish
• Task: Automatic Speech Recognition

Usage

from transformers import pipeline

# Load the model
asr = pipeline(
    "automatic-speech-recognition",
    model="northtechai/northstt-stageb-300",
    generate_kwargs={
        "condition_on_prev_tokens": True,
        "no_repeat_ngram_size": 2,
        "temperature": 0.0,
        "do_sample": False,
        "language": "tr",
    }
)

# Transcribe audio
result = asr("path/to/audio.wav")
print(result["text"])

Training Data

Training data information not provided.

Performance

This model has been fine-tuned for improved performance on Turkish speech recognition tasks.

Limitations

• Optimized for Turkish language
• Performance may vary on other languages
• Requires appropriate audio preprocessing (16kHz sample rate)

Citation

If you use this model, please cite the original Whisper paper:

@misc{radford2022robust,
  title={Robust Speech Recognition via Large-Scale Weak Supervision},
  author={Radford, Alec and Kim, Jong Wook and Xu, Tao and Brockman, Greg and McLeavey, Christine and Sutskever, Ilya},
  year={2022},
  eprint={2212.04356},
  archivePrefix={arXiv},
  primaryClass={cs.CL}
}