Hansong

Update setup.sh for tokenizer selection

4

Summary: For LLAMA3, users need to use tiktoken. Add a option to load from env var. Differential Revision: D56374673

[TEST ONLY] My release 0.2

1

Summary: * Update tutorial due to recent changes. * Clean up setup.sh for app helper lib build. Pull Request resolved: https://github.com/pytorch/executorch/pull/2962 Reviewed By: cccclai Differential Revision: D55951189 Pulled By: kirklandsign...

Reland Add more prebuilt artifacts

5

Summary: Build for different ABI in prebuild. Differential Revision: D56530383

Add support of mutable buffers in ExecuTorch

2

Context: https://github.com/pytorch/executorch/blob/main/examples/models/llama2/README.md?plain=1#L23 The current cache implementation in the original Llama2 repo is not supported by ExecuTorch, because ExecuTorch runtime assumes model data attributes being static.

Is it possible to build it for Android

4

It would be nice if we can try this out on Android. I followed instructions from https://github.com/ggerganov/ggml I tried it and seems that we need to remove `-march=native` from CMakeLists.txt...

Improve cmake testing setup

5

* Add build/Test.cmake and provide helper function for users to add tests targets Example: ``` et_cxx_test(my_test SOURCES my_test.cpp EXTRA_LIBS portable_kernels) ``` User need to invoke et_cxx_test and add their srcs/deps....