we introduce TimesFM, a single forecasting model pre-trained on a large time-series corpus of 100 billion real world time-points. Compared to the latest large language models (LLMs), TimesFM is much smaller (200M parameters), yet we show that even at such scales, its zero-shot performance on a variety of unseen datasets of different domains and temporal granularities come close to the state-of-the-art supervised approaches trained explicitly on these datasets.

Very impressive. I should try this on a couple of datasets.

see: https://github.com/google-research/timesfm

see: https://arxiv.org/abs/2310.10688

Generate optimized zig code at compile time to do neural network inference.

A machine learning primer built from first principles. For engineers who want to reason about ML systems the way they reason about software systems. - dreddnafious/thereisnospoon

a blog about machine learning, data, math, and other topics

Build your own ML framework. TinyTorch is organized into four progressive tiers that take you from mathematical foundations to production-ready systems. Each tier builds on the previous one, teaching you not just how to code ML components, but how they work together as a complete system.

Machine Learning Systems provides a systematic framework for understanding and engineering machine learning (ML) systems. This textbook bridges the gap between theoretical foundations and practical engineering, emphasizing the systems perspective required to build effective AI solutions. Unlike resources that focus primarily on algorithms and model architectures, this book highlights the broader context in which ML systems operate, including data engineering, model optimization, hardware-aware training, and inference acceleration. Readers will develop the ability to reason about ML system architectures and apply enduring engineering principles for building flexible, efficient, and robust machine learning systems.

Note from me, Frederick: I did not write this, but it reads exactly as if I had wrote it. Holy cow.

Imagine loading any track in Mixxx and instantly isolating the vocals, drums, bass, or instruments live, in real time. This is the vision behind our Google Summer of Code 2025 project: “Converting Demucs v4 (Hybrid Transformer) AI model to ONNX format”.

Will be a huge game-changer for DJing; very much looking forward to this.

The Rules of Machine Learning guide provides a set of best practices and distilled wisdom from Google engineers for building, deploying, and maintaining robust and effective Machine Learning systems in production.

This prophetic Bob the Angry Flower cartoon from 2003.

(SLOGOR)