Runtime implementation suggestion
Describe model format, tokenizer behavior, tensor layout, backend, platform, API contract, reproduction, expected parity, performance evidence, and the smallest testable change.
Use the LMRuntime packet →LMRuntime.com / Name and project boundaries
Runtime Disambiguation
A source-routed comparison of LMRuntime.com, Google LiteRT-LM, LM Studio inference runtimes, LinkMove LmRuntime, and generic language-model runtime terminology.
Name collision
“LM runtime” can refer to unrelated products and codebases.
Before proposing a feature, identify the exact project, model container, language, platform, and source authority. Similar names do not imply shared architecture, governance, compatibility, ownership, or contribution routes.
- This siteLMRuntime.com documents the published UAIX.LmRuntime local-first C# package family.
- Other projectsGoogle LiteRT-LM, LM Studio runtimes, and LinkMove LmRuntime have their own owners, formats, APIs, and feedback channels.
- Generic usage“Language-model runtime” can also describe any engine that loads and executes model artifacts.
Source-routed comparison
Which runtime do you mean?
This table is intentionally high level. Verify current details in each project’s official documentation before relying on them.
| Term or project | What it refers to | Typical artifact or interface | Where suggestions belong |
|---|---|---|---|
| LMRuntime.com / UAIX.LmRuntime | A published local-first C# package family for GGUF intake, tokenization, tensors, managed CPU kernels, LLaMA-family sessions, sampling, and a bounded application facade. | NuGet packages, C# APIs, GGUF-oriented runtime layers, package guides, coding examples, and a generated member reference. | LMRuntime idea guidance and the official contact path. |
| Google LiteRT-LM | Google AI Edge’s cross-platform orchestration layer for running LLMs with LiteRT, including hardware acceleration and multi-platform APIs. | .litertlm containers, CLI, and platform APIs. | Google AI Edge’s official LiteRT-LM source and documentation. |
| LM Studio inference runtimes | Inference backends managed by LM Studio and selectable from its application or CLI. | lms runtime commands for listing, downloading, selecting, removing, or updating runtimes. | LM Studio’s official runtime CLI documentation and its current project feedback channels. |
LinkMove LmRuntime | A Java runtime class in the LinkMove model-driven data acquisition and ETL framework. | Java APIs and LinkMove task/configuration model. | The official nhl/link-move source repository and its current contribution guidance. |
| Generic “LM runtime” | A broad description for software that loads language-model artifacts, manages execution state, and interfaces with hardware or APIs. | Varies: GGUF, proprietary packages, framework graphs, local services, libraries, or CLIs. | The owner of the concrete runtime being discussed. |
Route by change type
Implementation suggestions and theory suggestions are different work.
They can inform each other without sharing source authority.
Teleodynamic theory or review-process suggestion
Describe the governance concept, source, expected review effect, resource lane, claim boundary, human-review trigger, and why it should—or should not—change LMRuntime policy.
Review the governance model →Before sending a comparison
Name the project and verify the official source.
Do not infer compatibility or contribution routes from a similar name. Include the exact official URL, runtime version, model artifact, platform, and the behavior being compared.