Tag
This post builds a complete, accelerated tensor library from scratch in C, explaining the concept of tensors and their operations.
Introduces the Contagion Tensor and Coupling Amplification Factor (CAF) as a baseline-referenced, unitless ratio to quantify output-distribution coupling in multi-agent LLM systems, validated with real-API experiments on DeepSeek-Chat and GPT-4o-mini.
The author shares progress on building a CPU-only tensor library in C, covering basics like add/mul, reduce, strides, and 2D matmul, along with insights from reading Arcee's technical blogs on foundation models.
Noam Shazeer describes a coding convention for naming tensors with dimension suffixes to improve code readability and sanity, used at Character.AI since 2022.
A beginner-friendly GitHub repository covering PyTorch fundamentals, including tensor initialization, operations, indexing, and reshaping, with over 900 stars.