mohammad-ghaderi/cat-dog-asm-cnn

A Convolutional Neural Network implemented entirely from scratch in x86-64 assembly using AVX-512, performing cat vs dog image classification without any ML frameworks or libraries.

45
/ 100
Emerging

Implements complete forward and backward propagation with configurable multi-layer architecture (Conv2D, MaxPooling, dense layers, ReLU/Sigmoid activations) entirely in NASM assembly using AVX-512 to vectorize 16 float32 operations per cycle. Achieves ~10× speedup over NumPy despite sacrificing specialization for generality, trains on 25,000 128×128 RGB images, and includes custom debugging techniques for SIMD tensor validation without relying on traditional debuggers.

168 stars.

No Package No Dependents
Maintenance 10 / 25
Adoption 10 / 25
Maturity 13 / 25
Community 12 / 25

How are scores calculated?

Stars

168

Forks

14

Language

Assembly

License

MIT

Last pushed

Feb 08, 2026

Commits (30d)

0

Get this data via API

curl "https://pt-edge.onrender.com/api/v1/quality/ml-frameworks/mohammad-ghaderi/cat-dog-asm-cnn"

Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.