Public
  1. Public

onnxruntime

Public

Commits

Jiajia QinGitHub
05c8dc9d1c9
Jiajia Qin authored and GitHub committed 7e0dd9d4338
[js/webgpu] Optimize Expand (#22752)

Use components = 4 if possible.

llama3.2-1B becomes 20 tokens/s from 18 tokens/s on my iGPUs.