Speech Recognition of Moroccan Arabic Dialect using the Model HuBERT (available on HuggingFace) -- 2
Budget: $30 – $250 USD
I need a programmer familiar with HuggingFace and Speech Recognition.
What you will provide:
- code for pre-training HuBERT on a sample unlabeled arabic dataset that I will provide.
- code for fine-tuning HuBERT on a sample labeled arabic dataset I will provide.
Skills:
- You can use HuggingFace if you are familiar with it.
- There is a FB github repo: fairseq, that has all the necessary code for pre-training and fine-tuning. You can just use that repo and make it work with the very small datasets I will provide.
Budget:
- Around 80$, Negotiable.
Timeline:
- The project itself will require you few hours to 2 days. However, you can start the project anytime in the following 3 weeks (negotiable).
What you will provide:
- code for pre-training HuBERT on a sample unlabeled arabic dataset that I will provide.
- code for fine-tuning HuBERT on a sample labeled arabic dataset I will provide.
Skills:
- You can use HuggingFace if you are familiar with it.
- There is a FB github repo: fairseq, that has all the necessary code for pre-training and fine-tuning. You can just use that repo and make it work with the very small datasets I will provide.
Budget:
- Around 80$, Negotiable.
Timeline:
- The project itself will require you few hours to 2 days. However, you can start the project anytime in the following 3 weeks (negotiable).