ulatekh c8eeb93a6a
whisper : suppress tokens with a regex (#1997)
* Allow a regular expression to describe tokens to suppress.

Example: --suppress-tokens-re "[,\.]|[ ]?[0-9]+" will suppress commas, periods, and numeric tokens.

Technique inspired by https://github.com/openai/whisper/discussions/1041

Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>

* Blind change to fix Java test.

---------

Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
2024-04-09 18:27:28 +03:00
..
2024-01-05 17:11:27 +02:00
2024-03-21 07:40:09 +02:00
2022-12-12 20:17:27 +02:00