using `-m build --sdist` instead of `setup.py sdist`

* pre-commit autoupdate && pre-commit run --all-files

* Black formatter needs a current version of Python

* Bugfix: Illogical "Avoid computing higher temperatures on no_speech"

Bugfix for https://github.com/openai/whisper/pull/1279

It's "silence" when decoding has failed due to `compression_ratio_threshold` too, when further down the code it's not "silence" anymore.

"Silence" should be only when decoding has failed due to `logprob_threshold`.

Like described there:
https://github.com/openai/whisper/blob/8bc8860694949db53c42ba47ddc23786c2e02a8b/whisper/transcribe.py#L421

And in code there:
https://github.com/openai/whisper/blob/8bc8860694949db53c42ba47ddc23786c2e02a8b/whisper/transcribe.py#L243-L251



* Fix if "logprob_threshold=None"

---------

Co-authored-by: Jong Wook Kim <jongwook@openai.com>

Default now uses Turbo instead of Small

* Add option to carry initial_prompt with the sliding window

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

* Prevent redundant initial_prompt_tokens

* Revert unnecessary .gitignore change

---------

Co-authored-by: Kittsil <kittsil@gmail.com>
Co-authored-by: Jong Wook Kim <jongwook@openai.com>

* allowing numpy 2 in tests

* allowing numpy 2 in tests

* using sdpa if available

* Update model.py

* pinning numpy<2 in tests

* pip install together

* pip install together

* Relax triton requirements for compatibility with pytorch 2.4 and newer

Similar to https://github.com/openai/whisper/pull/1802, but now when pytorch upgrades to 2.4, it requires triton==3.0.0. I am not sure if it makes sense to remove the upper bound version constraints

* Update requirements.txt

* Add clip_timestamps option

* Add hallucination_silence_threshold option

* Fix typing for python < 3.9

---------

Co-authored-by: Jong Wook Kim <jongwook@openai.com>

* mel_filters() loads 128 mel bins

* can load 100-language models

* large-v3 checkpoint and evals

* add mandarin alias

* remove unused path

* flake8 fix

* formatting fix

* docs: defines relative speed in README

* combined paragraphs

---------

Co-authored-by: Jong Wook Kim <jongwook@nyu.edu>

* Update audio.py

 The `mel_filters` function is using a `np.load` function to load a pre-computed mel filterbank matrix. This function is not thread-safe, which means that if it is called from multiple threads at the same time, it may corrupt the data.

To fix this, you can use the `torch.load` function instead. This function is thread-safe, so it will not corrupt the data if it is called from multiple threads at the same time.

* Update audio.py

updated the docstring

* allow_pickle=False

* newline

---------

Co-authored-by: Jong Wook Kim <jongwook@nyu.edu>
Co-authored-by: Jong Wook Kim <jongwook@openai.com>

* handling transcribe() exceptions.

* printing stacktrace

---------

Co-authored-by: invalid <invalid@email.com>
Co-authored-by: Jong Wook Kim <jongwook@nyu.edu>
Co-authored-by: Jong Wook Kim <jongwook@openai.com>

* ADD parser for new argument --max_words_count

* ADD max_words_count in words_options
ADD warning for max_line_width compatibility

* ADD logic for max_words_count

* rename to max_words_per_line

* make them kwargs

* allow specifying file path by --model

* black formatting

---------

Co-authored-by: Jong Wook Kim <jongwook@nyu.edu>

Co-authored-by: Jong Wook Kim <jongwook@openai.com>

* Add .pre-commit-config.yaml

Co-authored-by: arthur <arthur@rtzr.ai>

* flake8 E741

---------

Co-authored-by: Jong Wook Kim <jongwook@openai.com>

Signed-off-by: haoshengqiang <haoshengqiang@xiaohongshu.com>
Co-authored-by: haoshengqiang <haoshengqiang@xiaohongshu.com>

fixed a few typos

* word timing tweaks

* comment on eot

* clearer comments

* avoid rearranging all kv_caches

* avoid calculating the same kv_cache from cross attn

* Update decoding.py

* linter fix

---------

Co-authored-by: Jong Wook Kim <jongwook@openai.com>

* Improve timestamp heuristics.

* Track pauses with last_speech_timestamp

prompt_reset_since is set before all_tokens is extended hence does not have the expected effect.