For AI agents: the complete documentation index is available at https://open-spaced-repetition.github.io/ts-fsrs/zh-CN/llms.txt, the full documentation bundle is available at https://open-spaced-repetition.github.io/ts-fsrs/zh-CN/llms-full.txt, and this page is available as Markdown at https://open-spaced-repetition.github.io/ts-fsrs/zh-CN/guide/model/fsrs-4dot5.md.
  • 简体中文
  • FSRS-4.5

    FSRS-4.5 有 17 个可训练权重。它修改了 FSRS-4 的遗忘曲线,同时保留相同的难度和成功回忆更新结构。

    公共符号和 Model 契约见模型概览。

    公式来源

    已依据 awesome-fsrs Wiki — FSRS-4.5 核对。

    遗忘曲线与间隔

    R(t,S)=(1+19t81S)−1/2R(t,S)=\left(1+\frac{19t}{81S}\right)^{-1/2}
    
    I(r,S)=81S19(r−2−1)I(r,S)=\frac{81S}{19}\left(r^{-2}-1\right)
    

    初始记忆

    S0(G)=clamp⁡(wG−1,Smin,Smax)S_0(G)=\operatorname{clamp}\left(w_{G-1},S_{min},S_{max}\right)
    
    D0(G)=clamp⁡(w4−(G−3)w5,Dmin,Dmax)D_0(G)=\operatorname{clamp}\left(w_4-(G-3)w_5,D_{min},D_{max}\right)
    

    难度

    D∗=D−w6(G−3)D^*=D-w_6(G-3)
    
    D′=clamp⁡(w7w4+(1−w7)D∗,Dmin,Dmax)D'=\operatorname{clamp}\left(w_7w_4+(1-w_7)D^*,D_{min},D_{max}\right)
    

    复习后的稳定性

    Hard 时 HG=w15H_G = w_{15},否则为 11;Easy 时 EG=w16E_G = w_{16},否则为 11。

    Sf=min⁡(w11D−w12((S+1)w13−1)ew14(1−R),S)S_f=\min\left(w_{11}D^{-w_{12}}\left((S+1)^{w_{13}}-1\right)e^{w_{14}(1-R)},S\right)
    
    Sr=S[1+ew8(11−D)S−w9(ew10(1−R)−1)HGEG]S_r=S\left[1+e^{w_8}(11-D)S^{-w_9}\left(e^{w_{10}(1-R)}-1\right)H_GE_G\right]
    
    Sraw′(G)={Sf,G=1Sr,G∈{2,3,4}S'_{raw}(G)=
    \begin{cases}
    S_f,&G=1\\
    S_r,&G\in\{2,3,4\}
    \end{cases}
    

    Again 分支会避免遗忘后的稳定性超过复习前的 SS。最后,模型通过 clamp 将 Sraw′S'_{\mathrm{raw}} 限制在 Smin⁡S_{\min} 到 Smax⁡S_{\max} 之间。