Transcription function cannot handle response format specifications of "text" or "vtt"

[matthew-at-qamcom]

Describe the bug

Transcription function cannot handle response format specifications of "text" or "vtt". When those are specified, an exception is raised.

To Reproduce

1. Using Python, obtain a transcript for an audio file: transcript = openai.Audio.transcribe("whisper-1", audio_file, response_format=response_format).
2. Note that response_format="json" and response_format="verbose_json" work without issue.
3. Replace response format with response_format="vtt" or response_format="text"
4. Note that the code now raises an exception json.decoder.JSONDecodeError

Code snippets

No response

OS

Linux

Python version

Python 3.10.6

Library version

openai 0.27.0