Full finetuning in float32 bit of whisper small on SADA 2022 Datset.
Model Details
Model Description
The model is trained on SADA 2022 datset for arabic transcription. I forced the decoder to use the arabic special tokens. This was finetuned from openai whisper small.