* deprecated `transcribe_stable()` for Faster-Whisper models and replaced it with `transcribe()` ([de0b42e](https://github.com/jianfch/stable-ts/commit/de0b42e301a015aba94d1fa3e6364e6787b1fd9e))
* removed redundant copying of encoder output in `DecodingTaskStable._get_audio_features()` ([9fefdb8](https://github.com/jianfch/stable-ts/commit/9fefdb8406be811674a9f34eb09df72a582c0a4e))
* refactored `non_whisper.py` into sub-package `non_whisper` ([e6c44f9](https://github.com/jianfch/stable-ts/commit/e6c44f97d0d56015e1a63134b61bed768153696e))
* refactored contents of `non_whisper.py` to `transcribe.py` in `non_whisper` ([e6c44f9](https://github.com/jianfch/stable-ts/commit/e6c44f97d0d56015e1a63134b61bed768153696e))
* refactored `align()` and `align_words()` to use `non_whisper.alignment.Aligner` ([2696a8b](https://github.com/jianfch/stable-ts/commit/2696a8b9ebabfd23fc7aad0520f4ef9d380a93c5))
* refactored `refine()` to use `non_whisper.refinement.Refiner` ([2696a8b](https://github.com/jianfch/stable-ts/commit/2696a8b9ebabfd23fc7aad0520f4ef9d380a93c5))
* added `adjust_gaps()` to `result.WhisperResult` ([4f7cff2](https://github.com/jianfch/stable-ts/commit/4f7cff235a1efe3c5e56ca5e6596c135c183fba8))
* added `convert_to_segment_level()` to `result.WhisperResult` ([4f7cff2](https://github.com/jianfch/stable-ts/commit/4f7cff235a1efe3c5e56ca5e6596c135c183fba8))
* added `custom_operation()` to `result.WhisperResult` ([08421e2](https://github.com/jianfch/stable-ts/commit/08421e2901da52a42031c603982e7526613938cd))
* added `record` to `remove_segment()` and `remove_word()` ([08421e2](https://github.com/jianfch/stable-ts/commit/08421e2901da52a42031c603982e7526613938cd))
* added `reassign_ids` to `WordTiming.add()`, `WordTiming.split()` ([08421e2](https://github.com/jianfch/stable-ts/commit/08421e2901da52a42031c603982e7526613938cd))
* added `clip_timestamps` to `transcribe()` ([9fefdb8](https://github.com/jianfch/stable-ts/commit/9fefdb8406be811674a9f34eb09df72a582c0a4e))
* added `align_words()` ([c176ecd](https://github.com/jianfch/stable-ts/commit/c176ecd888053eea8cb2a69a2d35981d0bc4b67a))
* added `ignore_special_periods` to `split_by_gap()`, `split_by_punctuation()`, `split_by_length()`, `split_by_duration()` ([1b00156](https://github.com/jianfch/stable-ts/commit/1b00156e04e0522af795501be34d0c9d63a8728a))
* added `ignore_special_periods()` to `result.WhisperResult` ([1b00156](https://github.com/jianfch/stable-ts/commit/1b00156e04e0522af795501be34d0c9d63a8728a))
* added `record` to `merge_all_segments()` ([1b00156](https://github.com/jianfch/stable-ts/commit/1b00156e04e0522af795501be34d0c9d63a8728a))
* added `alignment.py` and `refinement.py` to `non_whisper` ([e6c44f9](https://github.com/jianfch/stable-ts/commit/e6c44f97d0d56015e1a63134b61bed768153696e))
* added `non_whisper.alignment.Aligner` ([e6c44f9](https://github.com/jianfch/stable-ts/commit/e6c44f97d0d56015e1a63134b61bed768153696e))
* added `non_whisper.refinement.Refiner` ([e6c44f9](https://github.com/jianfch/stable-ts/commit/e6c44f97d0d56015e1a63134b61bed768153696e))
* added `options.py` ([e6c44f9](https://github.com/jianfch/stable-ts/commit/e6c44f97d0d56015e1a63134b61bed768153696e))
* added `refine()` to Faster-Whisper model instances [[de0b42e](https://github.com/jianfch/stable-ts/commit/de0b42e301a015aba94d1fa3e6364e6787b1fd9e)]
* fixed `vad=True` memory leak ([4711a01](https://github.com/jianfch/stable-ts/commit/4711a01a1af090985a48ce8ae0a29764d776f479))
* fixed `AudioLoader` AttributeError for `_denoised_save_path` ([9fefdb8](https://github.com/jianfch/stable-ts/commit/9fefdb8406be811674a9f34eb09df72a582c0a4e))
* fixed `align()` compatibility issue with new Faster-Whisper versions ([fe00aaf](https://github.com/jianfch/stable-ts/commit/fe00aaf2bd64603decb4c01557b1d39431809793), [c176ecd](https://github.com/jianfch/stable-ts/commit/c176ecd888053eea8cb2a69a2d35981d0bc4b67a))
* fixed `merge_all_segments()` redundant records in `regroup_history ([1b00156](https://github.com/jianfch/stable-ts/commit/1b00156e04e0522af795501be34d0c9d63a8728a))
* fixed `denoiser='noisereduce'` producing audio tensor with nan values ([852b39c](https://github.com/jianfch/stable-ts/commit/852b39cd2d41cd7e8946ceeb0716821a77ef4333))
* fixed `ignore_special_periods` to filter with the actual word instead of `WordTIming.__repr__` ([39e503d](https://github.com/jianfch/stable-ts/commit/39e503d90e6bab7ac3366c155ca716b97e2692c7))
* fixed `align()` unable to initialize tokenizer when `text` is result from faster-whisper model and `language` is unspecified ([25d9e13](https://github.com/jianfch/stable-ts/commit/25d9e136054afc0c303a98b35d797e29dd8284a0))
* fixed `transcribe()` causing `TqdmWarning: clamping frac to range [0, 1]` ([10c26b1](https://github.com/jianfch/stable-ts/commit/10c26b133228d6d84c11a1774be1bfb25906be63))
* changed `progress_callback` for `transcribe()` and `align()` to pass positional arguments ([9fefdb8](https://github.com/jianfch/stable-ts/commit/9fefdb8406be811674a9f34eb09df72a582c0a4e), [c176ecd](https://github.com/jianfch/stable-ts/commit/c176ecd888053eea8cb2a69a2d35981d0bc4b67a))
* changed default regroup algorithm ([1b00156](https://github.com/jianfch/stable-ts/commit/1b00156e04e0522af795501be34d0c9d63a8728a), [39e503d](https://github.com/jianfch/stable-ts/commit/39e503d90e6bab7ac3366c155ca716b97e2692c7))
* changed `even_split = True` to not ignore locked words ([1b00156](https://github.com/jianfch/stable-ts/commit/1b00156e04e0522af795501be34d0c9d63a8728a))
* changed `clamp_max()` to only ignore segments with less than 2 words ([39e503d](https://github.com/jianfch/stable-ts/commit/39e503d90e6bab7ac3366c155ca716b97e2692c7))
* updated `WhisperResult.lock()` to accept empty string `startswith` and `endswith` ([4f7cff2](https://github.com/jianfch/stable-ts/commit/4f7cff235a1efe3c5e56ca5e6596c135c183fba8))
* updated README.md ([4f7cff2](https://github.com/jianfch/stable-ts/commit/4f7cff235a1efe3c5e56ca5e6596c135c183fba8), [7dab171](https://github.com/jianfch/stable-ts/commit/7dab1714c5915e3efe070ae697ade1d04f5c8c42), [9fe1bf5](https://github.com/jianfch/stable-ts/commit/9fe1bf511862dccb669ff27f5fae9ae206b91a10), [c176ecd](https://github.com/jianfch/stable-ts/commit/c176ecd888053eea8cb2a69a2d35981d0bc4b67a), [de0b42e](https://github.com/jianfch/stable-ts/commit/de0b42e301a015aba94d1fa3e6364e6787b1fd9e))
* updated non-regrouping chainable methods to always save to `regroup_history` ([08421e2](https://github.com/jianfch/stable-ts/commit/08421e2901da52a42031c603982e7526613938cd))
* updated `AudioLoader` with ability to load only specific portions of the audio source ([9fefdb8](https://github.com/jianfch/stable-ts/commit/9fefdb8406be811674a9f34eb09df72a582c0a4e))
* updated `AudioLoader` to provide more informative error messages when yt-dlp fails to load an URL ([9fefdb8](https://github.com/jianfch/stable-ts/commit/9fefdb8406be811674a9f34eb09df72a582c0a4e))
* updated `refine()` to automatically add missing word-timestamps ([c176ecd](https://github.com/jianfch/stable-ts/commit/c176ecd888053eea8cb2a69a2d35981d0bc4b67a))
* updated `refine()` to automatically encode text with missing tokens in `result` ([c176ecd](https://github.com/jianfch/stable-ts/commit/c176ecd888053eea8cb2a69a2d35981d0bc4b67a))
* updated `ignore_special_periods` to not consider words that end with ".." as special ([39e503d](https://github.com/jianfch/stable-ts/commit/39e503d90e6bab7ac3366c155ca716b97e2692c7))
* updated `align()` to support `presplit` for Faster-Whisper models ([2696a8b](https://github.com/jianfch/stable-ts/commit/2696a8b9ebabfd23fc7aad0520f4ef9d380a93c5))
* updated `refine()` to support Faster-Whisper models; note that `refine()` will slower on Faster-Whisper models ([2696a8b](https://github.com/jianfch/stable-ts/commit/2696a8b9ebabfd23fc7aad0520f4ef9d380a93c5))