Repository navigation

feat(core): move to ryzenai-corelib 0.33.1, run every corelib model behind one engine, prefer AMD Q4 GGUFs, rename rai to core - #774

Merged

tawei-amd merged 10 commits into

chizamd:feature/dev01-update

Oct 16, 2026

Merged

feat(core): move to ryzenai-corelib 0.33.1, run every corelib model behind one engine, prefer AMD Q4 GGUFs, rename rai to core#774
tawei-amd merged 10 commits into
ROCm:mainfrom
chizamd:feature/dev01-update

Commits on Oct 16, 2026
- feat(rai): move to ryzenai-corelib 0.26, add Qwen3.6-35B-A3B, run Gemma 4 at 16k
  
  chizamd
  and
  cursoragent
  committed
  ae970e0
- refactor(rai): run Qwen3, Qwen3.5 and Gemma 4 as corelib's own models
  
  chizamd
  and
  cursoragent
  committed
  1005b2e
- feat(rai): move to ryzenai-corelib 0.28, run Qwen3.5 at 32k, align prefill chunks to pages
  
  chizamd
  and
  cursoragent
  committed
  ef82ce3
- feat(rai): move to ryzenai-corelib 0.29.2
  
  chizamd
  and
  cursoragent
  committed
  bb2f975
- feat(rai): move to ryzenai-corelib 0.29.4
  
  chizamd
  and
  cursoragent
  committed
  66ed175
- refactor: rename the rai backend to core and aie_next to dev04
  
  chizamd
  and
  cursoragent
  committed
  76d270e
- feat(core): move to ryzenai-corelib 0.29.7
  
  chizamd
  and
  cursoragent
  committed
  e35abd3
- feat(pull): prefer AMD Q4 GGUFs for core models, falling back to public Q8_0
  chizamd
  committed
  336a3d2
- refactor(phi4): run Phi-4-mini on corelib's phi4mini model, Q8_0 or Q4_G128
  chizamd
  committed
  7c97540
- feat(core): move to ryzenai-corelib 0.33.1
  chizamd
  committed
  bb91d58