Add option to carry initial_prompt with the sliding window

[kittsil]

Background
Whisper's transcribe() struggles with contextual proper nouns if they appear after the initial prompt has been consumed; see some experimental results here. This solves that issue by allowing the initial "context" prompt to be carried as the sliding window moves through the audio.

Changes
Add an option carry_initial_prompt = False to whisper.transcribe().

When carry_initial_prompt is set to True, initial_prompt is prepended to each internal decode() call's prompt. If there is not enough context space at the start of the prompt, the prompt is left-sliced to make space.

[kittsil]

There are outstanding issues with this PR:

1. I have not found the definition of the 224 context token length.
2. It prepends the initial_prompt to itself before enough tokens have been generated, resulting in a predilection toward looping.
3. I have not written tests.

Closing this PR since I can't find a way to move it to draft.

[ryanheise]

Closing this PR since I can't find a way to move it to draft.

How to: https://docs.github.com/en/pull-requests/collaborating-with-pull-requests/proposing-changes-to-your-work-with-pull-requests/changing-the-stage-of-a-pull-request

[ryanheise]

Also a relevant discussion here: #1040 (comment)

I have not found the definition of the 224 context token length.

It's part of the model dimensions itself, actually 448 tokens total, and half that for the prompt. The logic is in decoding.py if you look for self.n_ctx: int = model.dims.n_text_ctx and look for the references to it.

[kittsil]

@ryanheise Thank you for your input; it was helpful. Do you mind providing any additional feedback?

---

Aside: I did find the left-slice in the code, and it turns out that the docs are wrong, as actually the maximum prompt length is 223!

Confirming with the medium.en model...

>>> medium = torch.load('/home/kittsil/.cache/whisper/medium.en.pt')
>>> medium['dims']
{'n_mels': 80, 'n_vocab': 51864, 'n_audio_ctx': 1500, 'n_audio_state': 1024, 'n_audio_head': 16, 'n_audio_layer': 24, 'n_text_ctx': 448, 'n_text_state': 1024, 'n_text_head': 16, 'n_text_layer': 24}
>>> medium['dims']['n_text_ctx'] // 2 - 1
223