Model reference · open weights
Arabic-Whisper-CodeSwitching-Edition is an open-weight audio or speech model from MohamedRashad. AxForge deploys and operates it for you on dedicated EU-owned hardware — with the licence handled where one is required.
Available as managed deployment — configured and operated for you on dedicated EU hardware, quoted per deployment.
What it is
| Released by | MohamedRashad |
|---|---|
| Type | Audio & music |
| Task | Speech→text |
| Parameters (lead) | 1.5B |
| Runs with | transformers |
| Released | 2024-07-06 |
| Popularity | 508 downloads / month |
| Licence | Open, with conditions |
About
This model is a fine-tuned version of Whisper Large v2 by OpenAI, trained on an Arabic-English-code-switching dataset.
The Arabic-Whisper-CodeSwitching-Edition is designed to handle Arabic audio with embedded English words. This model enhances the original Whisper Large v2 by improving its performance on Arabic-English code-switching speech
The model can be used directly for transcribing Arabic speech that includes English words. It is particularly useful in multilingual environments where code-switching is common.
The model may not perform well on monolingual speech in languages other than Arabic or English, or on speech with code-switching in languages other than Arabic and English.
Users (both direct and downstream) should be made aware of the risks, biases, and limitations of the model. More information needed for further recommendations.
Use the code below to get started with the model.
from transformers import WhisperForConditionalGeneration, WhisperProcessor
processor = WhisperProcessor.from_pretrained("MohamedRashad/Arabic-Whisper-CodeSwitching-Edition")
model = WhisperForConditionalGeneration.from_pretrained("MohamedRashad/Arabic-Whisper-CodeSwitching-Edition")
# Example usage
inputs = processor("path_to_audio_file.wav", return_tensors="pt")
generated_ids = model.generate(inputs["input_features"])
transcription = processor.batch_decode(generated_ids, skip_special_tokens=True)
print(transcription)
@misc{rashad2024arabicwhisper,
title={Arabic-Whisper-CodeSwitching-Edition},
author={Mohamed Rashad},
year={2024},
url={https://huggingface.co/spaces/MohamedRashad/Arabic-Whisper-CodeSwitching-Edition},
}
Rashad, M. (2024). Arabic-Whisper-CodeSwitching-Edition. Retrieved from https://huggingface.co/spaces/MohamedRashad/Arabic-Whisper-CodeSwitching-Edition
From the published model card. Full card on the HuggingFace links in the sidebar.
Using it via the API
Once AxForge deploys arabic-whisper-codeswitching-edition for you, it answers on the OpenAI-compatible API — the same base URL and keys as every other model. (arabic-whisper-codeswitching-edition below is illustrative; you get the exact model name on deployment.)
$ curl -sS https://api.axforge.ai/v1/audio/transcriptions \ -H "Authorization: Bearer $AXFORGE_API_KEY" \ -F model="arabic-whisper-codeswitching-edition" -F file=@audio.mp3
Create an account — your API key is available in the console. 3M free tokens every 30 days with every new account.