Hyperparameter optimization - Wikipedia
6
公开标注数
6
参与人数
2026-07-24 09:41:30
首次 Whisper
本页的公开 Whisper
划选高亮2026-07-24 16:05:30
原文高亮摘录
“Population Based Training (PBT) learns both hyperparameter values and network weights”
Whisper 随想笔记
So it's basically evolution for tuning, pretty clever trick.
划选高亮2026-07-24 13:02:30
原文高亮摘录
“Grid search suffers from the curse of dimensionality, but is often embarrassingly parallel”
Whisper 随想笔记
Ha, "embarrassingly parallel" is the only thing saving grid search for me.
划选高亮2026-07-24 12:53:30
原文高亮摘录
“Grid search suffers from the curse of dimensionality, but is often embarrassingly parallel”
Whisper 随想笔记
True, but even parallel, it's still brute force if your grid's huge.
划选高亮2026-07-24 09:59:30
原文高亮摘录
“Hyperparameter optimization determines the set of hyperparameters that yields an optimal model”
Whisper 随想笔记
I always just guess and hope, but this makes me want to learn more.
划选高亮2026-07-24 09:50:30
原文高亮摘录
“Hyperparameter optimization determines the set of hyperparameters that yields an optimal model”
Whisper 随想笔记
Minimizing loss sounds easy until you actually try it.
划选高亮2026-07-24 09:41:30
原文高亮摘录
“Hyperparameter optimization determines the set of hyperparameters that yields an optimal model”
Whisper 随想笔记
That's the whole game really—tune or lose.
分享本页 Whisper
短链接
https://domwhisper.com/s/5fca71ff1bf0嵌入代码
<iframe src="https://domwhisper.com/embed/5fca71ff1bf0" width="100%" height="480" style="border:0;border-radius:16px" loading="lazy"></iframe>