Skip to content

v0.4.0

Choose a tag to compare

View all tags
@github-actions github-actions released this 08 Jul 18:41
v0.4.0
a2cbb31

Changes since v0.3.0:

Performance

  • 35B spec K=3 — 249/232/192, p1+p2 at margin (1.16x/1.11x); spec board 8/9 above llama
  • BW24_SPEC_M2 small-m batched verify (t=2 linear arm + MoE dev-rows), default OFF

Configuration

  • BW24_SPEC_M2 default ON — verify m-curve flattened (launch structure: t=2 batched arm + MoE dev-rows), 35B spec optimum shifts to K=3: 249/232/192 vs llama 215/208.5/201.7

Documentation

  • Hy3 REAP75 assessed as the next spilled-MoE target (42GB fits VRAM+RAM, no disk tier) — replaces M3's roadmap slot

Boards + reproduction artifacts: https://huggingface.co/Avifenesh/bw24-bench · full experiment log in research/tune-data/

Assets 2
Loading