正在加载视频...

视频加载失败

Implemented a visualization for the neural network today. This is a neural network being trained to learn the snake game in a terminal. It's built in Rust using Ratatui More input details in thread

165,655 次观看 • 2 年前 •via X (Twitter)

10 条评论

Bones 的头像
Bones2 年前

The newer arch has 24 inputs, 16 (8x2 directional vision) +. 8 (4x2, 1-hot encoded head and tail dirs) S - solid, F - food, LF/L - left, RT/R - right, BT/B - bottom, TP/T - top, TL - top left, TR - top right, BR - bottom right, BL - bottom left, HE - head, TA - tail.

Bones 的头像
Bones1 年前

Source:

mitch 🏴‍☠️ 的头像
mitch 🏴‍☠️2 年前

@ratatui_rs Literally on my to do list this week was developing my first neural network in rust, you’re the bomb

Sarlloc ✭ 的头像
Sarlloc ✭1 年前

@ratatui_rs how about this visualization?

Bones 的头像
Bones1 年前

@ratatui_rs This is really cool

Arnie Ramesh 的头像
Arnie Ramesh2 年前

@ratatui_rs This is sick

SifoDyas 的头像
SifoDyas2 年前

@ratatui_rs NEAT?

Bones 的头像
Bones2 年前

@ratatui_rs Yes, but this doesn't evolve the network architecture, just the weights/biases

Quentin Munch 的头像
Quentin Munch2 年前

@ratatui_rs @BonesaiDev what Rust algebra / NN lib did you use in Rust ? Or is it from scratch ?

Bones 的头像
Bones2 年前

@ratatui_rs Yup, everything was built from scratch, this is mostly a rewrite of own project -

相关视频