From Numerical Differences to Bitwise Prediction in Core ML ArithmeticHow targeted probes recover Core ML arithmetic rules and predict outputs bit for bit across tested CPU, GPU, and Neural Engine paths.77 min read·selfenrichment.hashnode.dev5d ago
Backend-Sensitive Numerical Differences and Discrete Decisions in Core MLThis is the full text of my paper, republished from Zenodo, with formatting adapted for the web. Robert Thomas Collins · Independent Researcher Abstract Core ML compute-unit choices can change individ43 min read·selfenrichment.hashnode.devOct 20
dgx-spark-inference - Keeping good habits around local inferenceLocal inference has a habit of becoming folklore. You launch a model from a shell history fragment you can never quite find again. A context setting lives in a note somewhere. A faster quantization ge8 min read·selfenrichment.hashnode.devJul 1
BF16 Edge Cases When Porting CUDA-Oriented TTS Models to MPS and MLX-AudioWhen porting CUDA-oriented TTS models to Apple Silicon, one subtle but important issue is BF16 behavior on Metal. This has shown up in DiaTTS work on MPS, and the same class of issue appears relevant8 min read·selfenrichment.hashnode.devJun 3
Revisiting MetalGraph: What I Learned About MLX, Replay, and Sidecar PerformanceTo set the stage, this post covers follow-up work that made MetalGraph feel less like it is doomed to be only a pure investigation tool and more like something that could be useful in practice. MetalG16 min read·selfenrichment.hashnode.devMay 11
