GitHub - ggml-org/llama.cpp: LLM inference in C/C++ · GitHub
5
公开标注数
5
参与人数
2026-07-16 09:47:54
首次 Whisper
本页的公开 Whisper
划选高亮2026-07-16 13:08:54
原文高亮摘录
“1.5-bit, 2-bit, 3-bit, 4-bit, 5-bit, 6-bit, and 8-bit integer quantization for faster inference and reduced memory use”
Whisper 随想笔记
I tried 4-bit on my laptop, runs way smoother than full precision.
划选高亮2026-07-16 12:59:54
原文高亮摘录
“1.5-bit, 2-bit, 3-bit, 4-bit, 5-bit, 6-bit, and 8-bit integer quantization for faster inference and reduced memory use”
Whisper 随想笔记
Damn, 1.5-bit sounds wild, does it even work or just a gimmick?
划选高亮2026-07-16 10:05:54
原文高亮摘录
“Plain C/C++ implementation without any dependencies”
Whisper 随想笔记
I tried to compile it on Windows and it was still a pain, but cool it's pure C.
划选高亮2026-07-16 09:56:54
原文高亮摘录
“Plain C/C++ implementation without any dependencies”
Whisper 随想笔记
Yeah, but no dependencies means you're on your own for everything else.
划选高亮2026-07-16 09:47:54
原文高亮摘录
“Plain C/C++ implementation without any dependencies”
Whisper 随想笔记
Finally something that doesn't need a bloated framework to run.
分享本页 Whisper
短链接
https://domwhisper.com/s/8f4d2163b5e6嵌入代码
<iframe src="https://domwhisper.com/embed/8f4d2163b5e6" width="100%" height="480" style="border:0;border-radius:16px" loading="lazy"></iframe>