search

second-state / WasmEdge-WASINN-examples

Apache License 2.0
217 stars 36 forks source link

[Example] Highlight how to use n_gpu_layers on CUDA hardware #58

Closed hydai closed 6 months ago

juntao commented 6 months ago

Hello, I am a code review bot on flows.network. Here are my reviews of code commits in this PR.

Commit f2cf66a74d656fbdf2d299b9f3add0192192b722

Key changes in the patch:

  1. Updated the installation command for WasmEdge to version 0.13.5.
  2. Added instructions for installing the ggml plugin on different platforms (macOS, CUDA-enabled Linux, CPU-only Linux).
  3. Added instructions for GPU acceleration using the n_gpu_layers parameter.
  4. Fixed some typos and errors in the README file.
  5. Added a new section for errors and troubleshooting.
  6. Added descriptions of supported parameters.

Potential problems:

  1. It is unclear why the installation command for WasmEdge is repeated multiple times in the patch.
  2. The patch mentions CUDA, but it is not clear if the instructions are specific to a certain version or distribution of CUDA.
  3. The patch does not provide a clear explanation of what the n_gpu_layers parameter does and how to choose an appropriate value.
  4. The instructions for GPU acceleration on macOS state that specifying n_gpu_layers is not necessary, but it is not clear why.
  5. The patch includes links to external resources, but it would be helpful to provide more information in the README file itself.