Piotr Wilkin (ilintar) ead6bf521b

fix: extend f32 matmul precision to ROCm for Qwen-Image, Krea2 and Boogu (#1772)

2026-07-10 23:23:01 +08:00

..

anima.hpp

feat: add qwen image layered support (#1119)

2026-07-03 00:21:22 +08:00

boogu.hpp

fix: extend f32 matmul precision to ROCm for Qwen-Image, Krea2 and Boogu (#1772)

2026-07-10 23:23:01 +08:00

control.hpp

fix: SDXL ControlNet (diffusers naming + graph size) (#1752)

2026-07-06 22:06:58 +08:00

dit.hpp

feat: add qwen image layered support (#1119)

2026-07-03 00:21:22 +08:00

ernie_image.hpp

sync: update ggml and revert vulkan workarounds for Anima and Ernie (#1710)

2026-06-27 00:52:32 +08:00

flux.hpp

refactor: move model-specific args into model parsers (#1757)

2026-07-06 23:13:18 +08:00

hidream_o1.hpp

chore: strip UTF-8 BOMs and add cleanup script (#1726)

2026-06-30 22:32:45 +08:00

ideogram4.hpp

refactor: route all runner params through model manager (#1649)

2026-06-14 02:05:23 +08:00

krea2.hpp

fix: extend f32 matmul precision to ROCm for Qwen-Image, Krea2 and Boogu (#1772)

2026-07-10 23:23:01 +08:00

lens.hpp

refactor: route all runner params through model manager (#1649)

2026-06-14 02:05:23 +08:00

ltxv.hpp

feat: support disk params backend (#1651)

2026-06-14 14:48:50 +08:00

minit2i.hpp

feat: add MiniT2I support (#1683)

2026-07-02 00:46:51 +08:00

mmdit.hpp

feat: support disk params backend (#1651)

2026-06-14 14:48:50 +08:00

model.hpp

feat: add qwen image layered support (#1119)

2026-07-03 00:21:22 +08:00

pid.hpp

refactor: route all runner params through model manager (#1649)

2026-06-14 02:05:23 +08:00

qwen_image.hpp

fix: extend f32 matmul precision to ROCm for Qwen-Image, Krea2 and Boogu (#1772)

2026-07-10 23:23:01 +08:00

sefi_image.hpp

chore: silence narrowing conversion warnings (#1717)

2026-06-28 23:14:53 +08:00

unet.hpp

refactor: route all runner params through model manager (#1649)

2026-06-14 02:05:23 +08:00

wan.hpp

feat: support disk params backend (#1651)

2026-06-14 14:48:50 +08:00

z_image.hpp

fix: avoid f16 overflow in Z-Image quantized matmuls on ROCm (#1771)

2026-07-10 23:22:15 +08:00