@nvt.210828:

T 🌊
T 🌊
Open In TikTok:
Region: VN
Monday 24 August 2026 10:17:09 GMT
600503
47190
86
4303

Music

Download

Comments

_dtmt_12
Trinh Đặng :
đúng ý
2026-08-24 10:58:14
60
hm.0311
m. :
Y với H đi ạ
2026-08-26 07:39:57
2
nguyentiendthwnhathegioi
TL. :
T với L đi ạ
2026-08-25 10:44:53
31
nho.cai.lon_
Dương Ý🙎 :
tuất trung🤗
2026-08-26 11:00:40
5
namet004
uyen trang. :
T với K dc kooo
2026-08-25 08:19:38
6
tnct_48
Ctuong kawaii😌 :
T vs Đ đi tus🥺🥺
2026-08-26 11:02:28
4
knockhah_n
nk. :
cứ tưởng =))
2026-08-26 21:30:03
2
hong.guen1
đcm 🤬 :
T vs H ik a
2026-10-03 03:57:18
1
sukaaa437
𝓑𝓲𝓲❤️‍🔥 :
k vs m ii
2026-08-27 05:38:44
2
_qnhu.27
qn :
toán 💗thi
2026-08-25 04:43:56
1
user6102485993178
Ha My volleyball :
M vs N ii ạ
2026-08-25 12:23:36
1
mai.tn59
PhướcTấnnn :
sớm r nè
2026-08-24 10:21:56
1
haybuon927_
🖤 :
sớm 💗💗💗
2026-08-24 10:19:24
1
chiec.lop.ben.bi2
HTK 💗 :
T vs C ikk tus
2026-08-27 04:44:26
1
lynh.081
𝐌𝐚𝐢 𝐋𝐢𝐧𝐡✨ :
H vs L điii ạ🥰
2026-08-26 04:18:05
1
todiddy4
H :
Tim vid cho mik vs ạ
2026-08-26 08:21:27
1
tng.ank61
T🌊 :
2026-08-25 00:32:35
1
nguyenthuu2100
nguyenthu10 :
2026-08-25 02:19:29
1
.thaouyen180612_
𝙩𝙝𝙖𝙤𝙤𝙪𝙮𝙚𝙣𝙣🎀 :
H vs U di dc hong ạ
2026-08-25 08:46:07
1
mjnhthuann
hăn :
1tỉ bước chân
2026-08-28 15:24:47
1
ba.khi._6a6
ᥫᩣThanh an :
s ko phải là T❤️A 😅
2026-10-10 06:27:23
1
kim.tin6987
Cua ngố :
Mới ct hihi
2026-10-16 07:27:49
0
soilamgi831
Nhi Tâm :
T và N
2026-11-08 07:12:16
0
quc.khnh10618
Cún :
CT rồi mà
2026-10-13 04:50:24
0
0711201h
6h1 🥀 :
đẹp
2026-09-04 10:15:09
0
To see more videos from user @nvt.210828, please go to the Tikwm homepage.

Other Videos


That it acts like a digital painter, sketching outlines and filling in colors. It doesn’t. Neural networks don't "think" in sequences, and they don't constantly ask themselves, "What should I draw next?" Instead, generating an image is mathematically identical to navigating through a dense fog. The AI is simply solving a differential equation, taking tiny steps through a multi-million-dimensional space until it finds the exact coordinate where your image exists. 🧠 **The Quick-Win Mnemonic: Remember the M.A.P.** Next time you use Midjourney or Stable Diffusion, remember that you are just setting up a M.A.P.: 📍 **M**ap = Your text prompt. It is encoded exactly *once* and acts as a constant radio beacon. 📍 **A**rea = The latent space. You always start dropped into a random coordinate of pure noise (the fog). 📍 **P**ath = The direction vector. The AI doesn't output a picture; it outputs a massive tensor of numbers telling you exactly which way to step to clear the fog. Different starting noise + Same constant map = A completely unique path to a new image. 🔬 **The Deep Dive (Link in Bio)** This carousel gives you the mental model, but the actual mathematics behind this—reversing entropy via Stochastic Differential Equations (SDEs) and Flow-Matching—is one of the most beautiful derivations in modern computer science. (And it’s the exact same math being used right now to discover new pharmaceutical drugs and fold proteins). I’ve written a full, academic-level breakdown of the math on Substack. Click the link in my bio to read "Sculpting from Static." 💬 **Question for you:** Did you previously think the AI was generating new, hidden prompts for itself at every step of the generation? Let me know below! 👇 #generativeai #stablediffusion #machinelearning #firstprinciples #stochasticcalculus

About