1. Linear Issues I’ve spent years building scaffolding for coding agents. Custom prompts, state machines, retry logic, context management systems. However, each time the infrastructure seemed to weigh the models down.
Then all of a sudden, they started actually using tools correctly. Turns out, they just needed to be trained on tool-use traces, and now the habits are baked in: when to call functions, how to recover from errors, when to retry versus escalate.
github pypi phase prediction per best epoch
phasefinder is a beat estimation model that predicts metric position as rotational phase, heavily inspired by this paper. Demos Your browser does not support the audio element. `are you looking up`
mk.gee Your browser does not support the audio element. `just the way it is`
action bronson Your browser does not support the audio element. `lethal weapon`
github pypi DeepRhythm is a convolutional neural network designed for rapid, precise tempo prediction for modern music. It runs on anything that supports Pytorch (I’ve tested Ubuntu, MacOS, Windows).
Audio is batch-processed using a vectorized harmonic constant-q modulation (HCQM), drastically reducing computation time by avoiding the usual bottlenecks encountered in feature extraction.
Why? I needed a way to accurately estimate tempo that was small & fast enough run on a Raspberry Pi.