From f58fa76bb6edffc0ce5f772198d3d17ef21b8176 Mon Sep 17 00:00:00 2001 From: Meinard Francisco <44516102+znarfm@users.noreply.github.com> Date: Sun, 23 Aug 2026 22:11:49 +0800 Subject: [PATCH] litert-lm: add page (#23600) Co-authored-by: Wiktor Perskawiec --- pages/common/litert-lm.md | 36 ++++++++++++++++++++++++++++++++++++ 1 file changed, 36 insertions(+) create mode 100644 pages/common/litert-lm.md diff --git a/pages/common/litert-lm.md b/pages/common/litert-lm.md new file mode 100644 index 0000000000..c213600905 --- /dev/null +++ b/pages/common/litert-lm.md @@ -0,0 +1,36 @@ +# litert-lm + +> Run and manage LiteRT-LM models. +> More information: . + +- List all imported models: + +`litert-lm list` + +- Run a local or imported model interactively: + +`litert-lm run {{path/to/model.litertlm|model_id}}` + +- Run a model directly from a Hugging Face repository with a single prompt: + +`litert-lm run --from-huggingface-repo {{owner/repository}} {{model.litertlm}} --prompt "{{What is the capital of France?}}"` + +- Run a model using GPU acceleration: + +`litert-lm run {{path/to/model.litertlm|model_id}} --backend gpu` + +- Run a multimodal model with an image attachment: + +`litert-lm run {{path/to/model.litertlm|model_id}} --vision-backend {{gpu|cpu}} --attachment {{path/to/image.jpg}} --prompt "{{Describe this image.}}"` + +- Run a model with Python function calling tools via a preset file: + +`litert-lm run {{path/to/model.litertlm|model_id}} --preset {{path/to/preset.py}}` + +- Benchmark a model's performance: + +`litert-lm benchmark {{path/to/model.litertlm|model_id}}` + +- Start an OpenAI-compatible API server on a specific port: + +`litert-lm serve --port {{9379}}`