You like pytorch? You like micrograd? You love tinygrad! ❤️
Repository Index
47,133 repositories indexed
1–19 of 19 · Most stars
A tiny scalar-valued autograd engine and a neural net library on top of it with PyTorch-like API
Suite of tools for deploying and training deep learning models using the JVM. Highlights include model import for keras, tensorflow, and onnx/pytorch, a modular and tiny c++ library for running math code and a java based math library on top of the core c++ library. Also includes samediff: a pytorch/tensorflow like library for running deep learn...
- java
- gpu
- deeplearning
- neural-nets
Minimal reproduction of DeepSeek R1-Zero
PyTorch implementation of YOLOv3, YOLOv3-SPP, and YOLOv3-tiny for real-time object detection with training, validation, inference, and multi-format export.
header only, dependency-free deep learning framework in C++14
- c-plus-plus
- deep-learning
- machine-learning
- neural-network
simple neural network library in C99
- backpropagation
- genetic-algorithm
- ann
- neural-network
A lightweight header-only library for using Keras (TensorFlow) models in C++.
Tiny YOLO for iOS implemented using CoreML but also using the new MPS graph API.
- core-ml
- mps
- metal
- machine-learning
🚧 Under construction — no-training RefMod reference adapters for MiniMax H3 (compress reference images/videos into tiny latents, no training needed).
- comfyui
- comfyui-custom-node
- comfyui-extension
- comfyui-manager
Translated斑海豹 (Spotted Seal) — Phocinae-Largha-150M-v1: a 144.3M bilingual decision model. Tiny model. Big decisions.
- bilingual
- calibration
- cpu-inference
- decision-making
Tiny world model generates games and stream ac plays them all on Pico 2W
- rust
- transformers
Whistle speech-to-text running fully on an ESP32-S3: hold a button, say a colour in one of 7 languages, the LED does it. No network.
- edge-ai
- esp32
- esp32-s3
- inmp441
Machine Learning and Digital Signal Processing for MicroPython
- embedded-systems
- micropython
- python
- tinyml
Legend of Elya — N64 game with a real 6.36M-parameter ternary transformer on the VR4300 MIPS III CPU. Zelda-style dungeon, AI NPCs, byte-level inference at 1.23 tok/s scalar / 2.19 tok/s on the RSP overlay (measured under ares, never on silicon). Built with libdragon.
TensorFlow Lite models for MIRNet for low-light image enhancement.
- tinyml
- tflite
- tensorflow2
- computer-vision
Address book for computer vision models.
- opencv
- open-source
- deep-neural-networks
- tensorflow
Machine learning compiler in Go. A from-scratch tinygrad port: graph-rewrite IR, autodiff as a compiler pass, zero-CGO WebGPU backend.
- autograd
- compiler
- deep-learning
- golang
Every AI coding-agent CLI (Claude Code, Codex, Gemini, Qwen, Vibe) reviews your diff in parallel, read-only, for one ship/no-ship verdict. One tiny script: no server, no MCP.