Guide
Benchmark methodology before results
A benchmark result is not publishable evidence until the model, runtime, hardware, workload, procedure, statistics, raw data, and limitations are recorded.
MiRust and GGUPF: documentation versus implementation
A clear project boundary prevents research pages, WordPress releases, and external runtime prototypes from inheriting one another’s support claims.
Standard ML is not a Small Language Model
The abbreviations SML and SLM refer to different concepts and must remain distinct in MiRust documentation.
Resource budgets as first-class architecture
Memory, time, power, storage, and confidence budgets should drive runtime decisions instead of appearing only after implementation.
Reading model profiles critically
A model profile is useful only when parameter count, artifact size, memory, context, format, precision, license, target, and verification are kept distinct.