- Standard ML (SML)
- A statically typed functional programming language. In MiRust content, SML never means Small Language Model.
- Small Language Model (SLM)
- A language model selected for a comparatively bounded resource or deployment envelope; the term does not imply a universal parameter threshold.
- Micro language model
- A practical deployment category for highly constrained or narrow workloads, evaluated by total resource behavior rather than parameter count alone.
- Parameter count
- The number of learned parameters; distinct from artifact size, runtime memory, and context size.
- Model file size
- The stored artifact size for a specific format and precision.
- Runtime memory
- Host or device memory required for weights, state, buffers, context, and implementation overhead.
- Context size
- The amount of input and generated state a model or runtime can process, using an explicitly stated unit.
- Backend
- A concrete execution implementation for a device, API, or operator set.
- Quantization
- A transformation that represents values with reduced precision under a named scheme and measured quality trade-off.
- Verification
- A recorded review of a source, version, environment, claim, or measurement at a UTC timestamp.
- MiRust.com
- The public project guide, research index, documentation, and evidence surface.
- GGUPF.MiRust.com
- The separately governed implementation project identified by the owner; its availability and claims require independent verification.