Whisper-small Fine-Tune for Somali language

Thanks **Sanchit** for your blog https://huggingface.co/blog/fine-tune-whisper#combine-to-create-a-whisperprocessor.
Using this fine-tuning example you provided, with my own custom Somali dataset I trained the data with Colab GPU for around 3.5 hours and get 13.23, as shown below.


I loaded the model with the runs into https://huggingface.co/MohamedUgas/whisper-small-so

My issue is when I tried to preload the fine-tuned model for the processor I got this error:

After running this:
from transformers import WhisperForConditionalGeneration, WhisperProcessor
model = WhisperForConditionalGeneration.from_pretrained("MohamedUgas/whisper-small-so")
processor = WhisperProcessor.from_pretrained("MohamedUgas/whisper-small-so")

I got this Error:

processor = WhisperProcessor.from_pretrained("MohamedUgas/whisper-small-so")
OSError: Can't load tokenizer for 'MohamedUgas/whisper-small-so'. If you were trying to load it from 'https://huggingface.co/models', make sure you don't have a local directory with the same name. Otherwise, make sure 'MohamedUgas/whisper-small-so' is the correct path to a directory containing all relevant files for a WhisperTokenizer tokenizer.

All the files are in https://huggingface.co/MohamedUgas/whisper-small-so including the Colab notebook jupiter code.

Thanks for your help.
Mohamed

Start with the linked Hugging Face model repository and its Colab Jupyter notebook, then reproduce the WhisperProcessor.from_pretrained call shown in the issue. Check the model repository contents against the files required for WhisperTokenizer; done means the fine-tuned model and processor load without the reported OSError.

Written by the indexing model from the issue text.