music-explorer
tinygrad
Our great sponsors
music-explorer | tinygrad | |
---|---|---|
5 | 58 | |
31 | 17,800 | |
- | - | |
8.0 | 9.7 | |
4 days ago | 10 months ago | |
Shell | Python | |
- | 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.
music-explorer
-
When do we stop finding new music?
The article might describe a common scenario, but there are plenty of outliers. I hardly listen to music I liked in my teens and early twenties. I love discovering new music.
Many comments here are very insightful and discuss phenomena like high music diversity, music proliferation and easy of producing music, and automated recommendations.
One thing that has been occupying me is that curation is still harder than I'd like when using streaming tools like Spotify, YouTube Music, Apple Music, Tidal. Pandora had good roots with its music genome project, and have built on that. (I can't use it without a VPN since they discontinued supporting the country I mostly live in). It's probably a function of how I consume my music today - no longer desk-bound at work, but on the go, so iPhone (and Apple Watch) are primary tools. Being able to select/skip/preview/tune what I'm listening to is nowhere near as powerful as I'd like. I've written library curation tools in the past, these always expected me to spend significant dedicated time in front of a screen (e.g. a similar tool like the cool looking https://github.com/kristopolous/music-explorer, I think).
This has strong parallels to how older people consumed music - either totally passive curation (radio), or very deliberate (find music in record stores, at a friend's place, and/or select records/CDs in your own shelves). Today's ephemeral digital libraries are much lower effort, are huge and curation/selection tools are not easy enough to use, so I tend to fall back onto old favourites or recommendation engines that usually don't satisfy me.
A solution would be a much more configurable curation assistant that is also super easy to use (and, in my case) very accessible on a mobile device with 0-1 clicks (because I'm busy doing other things).
-
Goodbye Spotify
Might as well drop what I use for my music discovery, my fairly poorly documented hacker-friendly set of tools:
https://github.com/kristopolous/music-explorer/
-
Write Posix Shell
I'm a big fan of not posix bit instead modern bash and to all the complainers about dash and ash, I say "tough cookies".
Sometimes I'll even use zsh
Here's some example of a modern tool I have written for a subject I call "music discovery"
https://github.com/kristopolous/music-explorer/tree/master/t...
You'll see many languages in there.
If you don't like my practice then I guess don't use it. I've been using/developing these particular tools nearly every day for over 3 years and it works well for me.
I'm not going to say bash is awesome but it's pretty great for programming.
I use zsh as my interactive though
-
Why DRY is the most over-rated programming principle
Sure. Related. It's an art.
Here's some code I wrote earlier, probably a good example
https://github.com/kristopolous/music-explorer/blob/master/w...
It's self contained, not very big, not trying to be fancy, as direct as possible
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?
FizzBuzzEnterpris
Pytorch - Tensors and Dynamic neural networks in Python with strong GPU acceleration
dehydrated - letsencrypt/acme client implemented as a shell-script – just add water
llama.cpp - LLM inference in C/C++
ShellCheck - ShellCheck, a static analysis tool for shell scripts
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.
FizzBuzz Enterprise Edition - FizzBuzz Enterprise Edition is a no-nonsense implementation of FizzBuzz made by serious businessmen for serious business purposes.
llama - Inference code for Llama models
bash-modules - Useful modules for bash
tensorflow_macos - TensorFlow for macOS 11.0+ accelerated using Apple's ML Compute framework.
GPTQ-for-LLaMa - 4 bits quantization of LLaMA using GPTQ
text-generation-webui - A Gradio web UI for Large Language Models. Supports transformers, GPTQ, AWQ, EXL2, llama.cpp (GGUF), Llama models.