正在加载视频...

视频加载失败

Andrej Karpathy on Tensor Cores and TF32 precision

112,983 次观看 • 1 年前 •via X (Twitter)

9 条评论

Tommy Falkowski 的头像
Tommy Falkowski1 年前

He's an awesome teacher! I just love his clean style of explaining things!

ℏεsam 的头像
ℏεsam1 年前

he really has it in him

Konstantin 的头像
Konstantin1 年前

Can you share a full source please?

rogue node 的头像
rogue node1 年前

So basically TF32 is a lie.

Jacob Posel 的头像
Jacob Posel1 年前

Where does he post this?

john 的头像
john1 年前

Lmao what are the odds that the moment I get to this part of the video on my first time watch, i see this on twitter right after

ℏεsam 的头像
ℏεsam1 年前

i'm watching you :)

Felix Chin MD 的头像
Felix Chin MD1 年前

TF32 just means ur neural net’s matmuls are running on 18-bit numerics (FP 18 precision) apple (mlx) easily lets u control this (by setting intermediate weights and values to FP 16) don’t see how TF32 is anything special. seems like a marketing gimmick

MarcoDagna e/acc 🇪🇸 ⌛️🌍 的头像
MarcoDagna e/acc 🇪🇸 ⌛️🌍1 年前

@HeyGenLabs english spanish

相关视频