News for hackers. Stuff that matters.
Re:AlphaEvolve: A Gemini-powered coding agent for designing ... by cavisne on Wednesday May 14, 2025 @10:42PM (#43991313)
From the paper it was a speedup on the XLA GPU kernel they wrote using Jax, which is probably not SOTA. I don't think Jax even has a official flash attention implementation.
Re:AlphaEvolve: A Gemini-powered coding agent for designing ... by yarri on Thursday May 15, 2025 @11:58AM (#43996374)
Not sure what “official” means but would direct you to the GCP MaxText [0] framework which is not what this GDM paper is referring to but rather this repo contains various attention implementations in MaxText/layers/attentions.py

[0] https://github.com/AI-Hypercomputer/maxtext