@@ -344,7 +344,34 @@ litert benchmark model.tflite --gcp --device "pixel 7" --gcp-project "your-gcp-p
344344litert benchmark model.tflite --gcp --devices " pixel 7, sm-s931u1" --gpu
345345```
346346
347- ### 7. Visualize a model's architecture
347+ ### 7. Run and benchmark a generative LLM model using LiteRT-LM CLI
348+
349+ ` litert lm ` command will utlitize ` litert-lm ` , and you can use the same command
350+ with ` litert-lm ` , for example, both ` litert lm run ` and ` litert-lm run ` or
351+ ` litert lm benchmark ` and ` litert-lm benchmark ` achieve the same results.
352+
353+ Please follow the
354+ [ LiteRT-LM CLI guide] ( https://ai.google.dev/edge/litert-lm/cli ) for detailed
355+ instructions.
356+
357+ ``` bash
358+ # Run a generative LLM model, and load from hugging face
359+ litert lm run \
360+ --from-huggingface-repo=litert-community/gemma-4-E2B-it-litert-lm \
361+ gemma-4-E2B-it.litertlm \
362+ --prompt=" What is the capital of France?"
363+
364+ # Or load from a local LLM model file
365+ litert lm run ./my_model.litertlm
366+
367+ # Example with a custom prompt
368+ litert lm run ./my_model.litertlm --prompt " Hello, how are you?"
369+
370+ # Benchmark a generative LLM model
371+ litert lm benchmark ./my_model.litertlm
372+ ```
373+
374+ ### 8. Visualize a model's architecture
348375
349376``` bash
350377# Open in Model Explorer graph
@@ -354,7 +381,7 @@ litert visualize model.tflite
354381litert visualize --stop-all
355382```
356383
357- ### 8 . Import a local model
384+ ### 9 . Import a local model
358385
359386``` bash
360387# Import a local file into the centralized cache
@@ -364,7 +391,7 @@ litert import my_model.tflite --model-ref my_model
364391litert import ./my_model_dir --model-ref my_model --hf-id my_org_name/my_model
365392```
366393
367- ### 9 . List managed models
394+ ### 10 . List managed models
368395
369396``` bash
370397# List all managed models
@@ -374,40 +401,12 @@ litert list
374401litert list my_model
375402```
376403
377- ### 10 . Delete a managed model
404+ ### 11 . Delete a managed model
378405
379406``` bash
380407# Delete a model from cache
381408litert delete my_model
382409```
383-
384- ### 11. Run and benchmark a generative LLM model using LiteRT-LM CLI
385-
386- ` litert lm ` command will utlitize ` litert-lm ` , and you can use the same command
387- with ` litert-lm ` , for example, both ` litert lm run ` and ` litert-lm run ` or
388- ` litert lm benchmark ` and ` litert-lm benchmark ` achieve the same results.
389-
390- Please follow the
391- [ LiteRT-LM CLI guide] ( https://ai.google.dev/edge/litert-lm/cli ) for detailed
392- instructions.
393-
394- ``` bash
395- # Run a generative LLM model, and load from hugging face
396- litert lm run \
397- --from-huggingface-repo=litert-community/gemma-4-E2B-it-litert-lm \
398- gemma-4-E2B-it.litertlm \
399- --prompt=" What is the capital of France?"
400-
401- # Or load from a local LLM model file
402- litert lm run ./my_model.litertlm
403-
404- # Example with a custom prompt
405- litert lm run ./my_model.litertlm --prompt " Hello, how are you?"
406-
407- # Benchmark a generative LLM model
408- litert lm benchmark ./my_model.litertlm
409- ```
410-
411410### 12. Clean up all caches
412411
413412``` bash
0 commit comments