micrograd
tinygrad
Our great sponsors
micrograd | tinygrad | |
---|---|---|
22 | 58 | |
8,273 | 17,800 | |
- | - | |
0.0 | 9.7 | |
5 days ago | 10 months ago | |
Jupyter Notebook | Python | |
MIT License | MIT License |
Stars - the number of stars that a project has on GitHub. Growth - month over month growth in stars.
Activity is a relative number indicating how actively a project is being developed. Recent commits have higher weight than older ones.
For example, an activity of 9.0 indicates that a project is amongst the top 10% of the most actively developed projects that we are tracking.
micrograd
-
Micrograd-CUDA: adapting Karpathy's tiny autodiff engine for GPU acceleration
I recently decided to turbo-teach myself basic cuda with a proper project. I really enjoyed Karpathy’s micrograd (https://github.com/karpathy/micrograd), so I extended it with cuda kernels and 2D tensor logic. It’s a bit longer than the original project, but it’s still very readable for anyone wanting to quickly learn about gpu acceleration in practice.
-
Stuff we figured out about AI in 2023
FOr inference, less than 1KLOC of pure, dependency-free C is enough (if you include the tokenizer and command line parsing)[1]. This was a non-obvious fact for me, in principle, you could run a modern LLM 20 years ago with just 1000 lines of code, assuming you're fine with things potentially taking days to run of course.
Training wouldn't be that much harder, Micrograd[2] is 200LOC of pure Python, 1000 lines would probably be enough for training an (extremely slow) LLM. By "extremely slow", I mean that a training run that normally takes hours could probably take dozens of years, but the results would, in principle, be the same.
If you were writing in C instead of Python and used something like Llama CPP's optimization tricks, you could probably get somewhat acceptable training performance in 2 or 3 KLOC. You'd still be off by one or two orders of magnitude when compared to a GPU cluster, but a lot better than naive, loopy Python.
[1] https://github.com/karpathy/llama2.c
[2] https://github.com/karpathy/micrograd
-
Writing a C compiler in 500 lines of Python
Perhaps they were thinking of https://github.com/karpathy/micrograd
- Linear Algebra for Programmers
- Understanding Automatic Differentiation in 30 lines of Python
-
Newbie question: Is there overloading of Haskell function signature?
I was (for fun) trying to recreate micrograd in Haskell. The ideia is simple:
-
[D] Backpropagation is not just the chain-rule, then what is it?
Check out this repo I found a few years back when I was looking into understanding pytorch better. It's basically a super tiny autodiff library that only works on scalars. The whole repo is under 200 lines of code, so you can pull up pycharm or whatever and step through the code and see how it all comes together. Or... you know. Just read it, it's not super complicated.
-
Neural Networks: Zero to Hero
I'm doing an ML apprenticeship [1] these weeks and Karpathy's videos are part of it. We've been deep down into them. I found them excellent. All concepts he illustrates are crystal clear in his mind (even though they are complicated concepts themselves) and that shows in his explanations.
Also, the way he builds up everything is magnificent. Starting from basic python classes, to derivatives and gradient descent, to micrograd [2] and then from a bigram counting model [3] to makemore [4] and nanoGPT [5]
[1]: https://www.foundersandcoders.com/ml
[2]: https://github.com/karpathy/micrograd
[3]: https://github.com/karpathy/randomfun/blob/master/lectures/m...
[4]: https://github.com/karpathy/makemore
[5]: https://github.com/karpathy/nanoGPT
-
Rustygrad - A tiny Autograd engine inspired by micrograd
Just published my first crate, rustygrad, a Rust implementation of Andrej Karpathy's micrograd!
-
Hey Rustaceans! Got a question? Ask here (10/2023)!
I've been trying to reimplement Karpathy's micrograd library in rust as a fun side project.
tinygrad
- tinygrad: extreme simplicity, easiest framework to add new accelerators to
-
GGML – AI at the Edge
Might be a silly question but is GGML a similar/competing library to George Hotz's tinygrad [0]?
[0] https://github.com/geohot/tinygrad
-
Render neural network into CUDA/HIP code
at first glance i thought may its like tinygrad. but looks has many ops than that tiny grad but most maps to underlying hardware provided ops?
i wonder how well tinygrad's apporach will work out, ops fusion sounds easy, just a walk a graph, pattern match it and lower to hardware provided ops?
Anyway if anyone wants to understand the philosophy behind tinygrad, this file is great start https://github.com/geohot/tinygrad/blob/master/docs/abstract...
-
llama.cpp now officially supports GPU acceleration.
There are currently at least 3 ways to run llama on m1 with GPU acceleration. - mlc-llm (pre-built, only 1 model has been ported) - tinygrad (very memory efficient, not that easy to integrate into other projects) - llama-mps (original llama codebase + llama adapter support)
- George Hotz building an AMD competitor to Nvidia.
-
George Hotz ROCm adventures
Hopefully we will see now full support with AMD hardware on https://github.com/geohot/tinygrad. You can read more about it on https://tinygrad.org/
-
The Coming of Local LLMs
tinygrad
https://github.com/geohot/tinygrad/tree/master/accel/ane
But I have not tested it on Linux since Asahi has not yet added support.
llama.cpp runs at 18ms per token (7B) and 200ms per token (65B) without quantization.
- Everything we know about Apple's Neural Engine
- Everything we know about the Apple Neural Engine (ANE)
- How 'Open' Is OpenAI, Really?
What are some alternatives?
deepnet - Educational deep learning library in plain Numpy.
Pytorch - Tensors and Dynamic neural networks in Python with strong GPU acceleration
deeplearning-notes - Notes for Deep Learning Specialization Courses led by Andrew Ng.
llama.cpp - LLM inference in C/C++
ML-From-Scratch - Machine Learning From Scratch. Bare bones NumPy implementations of machine learning models and algorithms with a focus on accessibility. Aims to cover everything from linear regression to deep learning.
openpilot - openpilot is an open source driver assistance system. openpilot performs the functions of Automated Lane Centering and Adaptive Cruise Control for 250+ supported car makes and models.
NNfSiX - Neural Networks from Scratch in various programming languages
llama - Inference code for Llama models
yolov7 - Implementation of paper - YOLOv7: Trainable bag-of-freebies sets new state-of-the-art for real-time object detectors
tensorflow_macos - TensorFlow for macOS 11.0+ accelerated using Apple's ML Compute framework.
machine.academy - Neural Network training library in C++ and C# with GPU acceleration
GPTQ-for-LLaMa - 4 bits quantization of LLaMA using GPTQ