meta-llama / llama-recipes

Scripts for fine-tuning Meta Llama3 with composable FSDP & PEFT methods to cover single/multi-node GPUs. Supports default & custom datasets for applications such as summarization and Q&A. Supporting a number of candid inference solutions such as HF TGI, VLLM for local or cloud deployment. Demo apps to showcase Meta Llama3 for WhatsApp & Messenger.
9.95k stars 1.4k forks source link

[MLC-LLM] Introducing Llama 3 running locally on Android using MLC-LLM #506

Closed tmoreau89 closed 2 weeks ago

tmoreau89 commented 2 weeks ago

What does this PR do?

This PR introduces a guide on how to run Llama 3 8B instruct on an Android Phone using MLC LLM!

Feature/Issue validation/testing

This tutorial was tested with the following setup:

Before submitting

Thanks for contributing 🎉!

tmoreau89 commented 2 weeks ago

CC @HamidShojanazeri - the long awaited MLC-LLM tutorial!

tmoreau89 commented 2 weeks ago

@HamidShojanazeri we're green on spellcheck here, but the linter has issues with the following two URLs:

The error saying Status: 0 Error: Exceeded maxRedirects. Probably stuck in a redirect loop.

Unclear why but is there a way to get the linter to ignore these URLs?

HamidShojanazeri commented 2 weeks ago

sorry @tmoreau89 its the linter bug, can you pls add <!-- markdown-link-check-disable --> ,<!-- markdown-link-check-enable --> surrounding the link, https://github.com/meta-llama/llama-recipes/pull/291/files

tmoreau89 commented 2 weeks ago

@HamidShojanazeri I think we're good to go on this one! Thanks for the tip!

HamidShojanazeri commented 2 weeks ago

@tmoreau89 seems its still failing on "./mlc-package-config.json → Status: 400 [Error: ENOENT: no such file or directory, access '/github/workspace/recipes/inference/mobile_inference/android_inference/mlc-package-config.json']", would appreciate the fix to make the CI green thanks!

tmoreau89 commented 2 weeks ago

@HamidShojanazeri Morning! The checks are all green now (this time for real 😆 )