無所不在的拉格朗日 - Lagrangian Everywhere
revision control
Math AI - Expand Score Matching to Flow Matching
Posted on
|
In
GenAI
Flow Matching 與 Score Matching 的理論基礎比較
Math AI - Diffusion Acceleration Phases
Posted on
|
In
GenAI
如何加速 Diffusion Process
Math AI - Expand Score Matching to Flow Matching
Posted on
|
In
GenAI
Flow $\phi_t$ = 連續的 (自己到自己) 座標變換
Flow $\phi_t$ 定義: $x_t \triangleq \phi_t\left(x_0\right)$, 顯然 $t=0 \to x_0 = \phi_0\left(x_0\right)$ $\phi_t(x)$ 必須可微分且可逆。
Math AI - Normalizing Flow
Posted on
|
In
GenAI
Normalizing Flow (NF) - (自己到自己) 坐標變換 + Jacobian
Math AI - Expand Score Matching to Flow Matching
Posted on
|
In
GenAI
Takeaway
AI Coach for Bouldering Project
Posted on
|
In
CV
Introduction
From RNN to Linear Attention to Mamba
Posted on
|
In
Science
Source
Autoregression vs. Diffusion
Posted on
|
In
GenAI
Generative AI 兩大流派：CV (Computer Vision) and NL (Natural Language)
Discrete Diffusion
Posted on
|
In
GenAI
Excellent lectures!!!!: https://www.youtube.com/watch?v=8mxCNMJ7dHM&list=PL0H3pMD88m8XPBlWoWGyal45MtnwKLSkQ