---
language:
- kal
license: mit
metrics:
- wer
model_name: Whisper Tiny Fine-tuned on Kalaallisut
pipeline_tag: automatic-speech-recognition
tags:
- whisper
- automatic-speech-recognition
- speech-to-text
- kalaallisut
- greenlandic
widget:
- src: path/to/sample_audio.mp3
---

# Whisper Tiny Fine-tuned on Kalaallisut

This model is a fine-tuned version of the [openai/whisper-tiny](https://huggingface.co./openai/whisper-tiny) on a very small dataset of the Kalaallisut (Greenlandic) language. Whisper is a general-purpose speech recognition model trained on a large-scale dataset. However, this fine-tuned version on Kalaallisut is **not fully reliable** due to the limited amount of training data.

### Model Details

- **Model**: Whisper Tiny Fine-tuned on Kalaallisut
- **Base Model**: [openai/whisper-tiny](https://huggingface.co./openai/whisper-tiny)
- **Training Data**: A very small amount of audio-transcription pairs in Kalaallisut
- **Purpose**: Speech-to-text for the Kalaallisut language
- **License**: [MIT License](https://opensource.org/licenses/MIT)

### Training Data

This model was fine-tuned on a small dataset of the Kalaallisut language. The dataset was **not comprehensive** and **does not cover all aspects** of the language. As a result, the model is not reliable for general use cases and may produce incorrect transcriptions.

### Limitations

- The model has been trained on a **very small dataset** (only a few hours of audio).
- The **transcriptions may not be accurate**, especially for more complex audio inputs.
- This model is a **proof of concept** and not intended for production use.

### Intended Use

The model can be used for:

- Experimentation and testing with Kalaallisut language speech-to-text tasks.
- It may be helpful for small projects or as a foundation for further fine-tuning with more data.

### Usage Example

You can use the model for transcription with the following code:

```python
from transformers import WhisperProcessor, WhisperForConditionalGeneration
import torch

# Load the model and processor
processor = WhisperProcessor.from_pretrained("VoiceLessQ/whisper-tiny-kalaallisut")
model = WhisperForConditionalGeneration.from_pretrained("VoiceLessQ/whisper-tiny-kalaallisut")

# Load and process an audio file
audio_array = ...  # Load your audio file here
input_features = processor(audio_array, sampling_rate=16000, return_tensors="pt").input_features

# Generate transcription
generated_ids = model.generate(input_features)
transcription = processor.batch_decode(generated_ids, skip_special_tokens=True)[0]

print(f"Transcription: {transcription}")