On May 15th, version 1.6.0 of the high-performance automatic speech recognition system was released. whisper.cpp, implementing the model Whisper by OpenAI in C and C++, and distributed under the MIT license.
Change log:
- A new option —flash-attn has been added for faster data processing on devices with CUDA and Metal;
- higher performance on the ppc64 architecture;
- a bug in the processing of regular expressions in the sample console utility main has been fixed.
Source: linux.org.ru
