新发现
lnenad/lipwise — 将语音引擎构建固定在 AVX2 基准
lnenad/lipwise 将其内置语音识别引擎固定在 AVX2·FMA·F16C·BMI2 基准上,而不是跟随构建服务器的 CPU,从而阻止了由具备 AVX-512 的运行机器所产出的 Windows 构建出现 illegal instruction 崩溃。
lnenad/lipwise 将其内置语音识别引擎固定在 AVX2·FMA·F16C·BMI2 基准上,而不是跟随构建服务器的 CPU,从而阻止了由具备 AVX-512 的运行机器所产出的 Windows 构建出现 illegal instruction 崩溃。
GitHub 项目说明: Private voice dictation for Windows, macOS and Linux. Hold a shortcut, talk, and Lipwise types what you meant into any app: speech recognition runs on your computer, and an AI editor understands "scratch that", lists and rewrites. Works with local AI, Claude or OpenAI.
在 GitHub 查看 ↗lnenad/lipwise 将其内置语音识别引擎固定在 AVX2·FMA·F16C·BMI2 基准上,而不是跟随构建服务器的 CPU,从而阻止了由具备 AVX-512 的运行机器所产出的 Windows 构建出现 illegal instruction 崩溃。
没有符合条件的动态。