[EAGLE-5211] Refract openai whisper example

[luv-bansal]

What

• Implemented stream generation of the output text from any audio files in the generate method i.e it will transcribe the the audio file in the stream way even though openai whisper API don't support streaming
• corrected both generate and stream methods for batch requests
• Refracted Openai whisper example, now the model.py code is more readable

How

We got the input audio in bytes, first split the original audio into chunks of audios, Each chunk is a valid WAV file with proper headers, allowing them to be processed independently, then run the whisper API for each audio WAV file also ensuring that batch requests also supported.

Test

Tested this model example locally using local dev runner