Marco Colombo 82324fc508 hexagon: accept the zeroed rope probe in supports_op (#28995) ...

llama probes weight placement with a rope where all params are 0, so rejecting
n_dims == 0 or freq_base == 0 puts rope_freqs on the CPU. That splits the decode
graph at every full-attention layer (gemma-4-E2B: 5 splits instead of 2).

Assisted-by: Claude Opus 5

2026-09-16 08:38:44 -07:00

..

cmake

CUDA: replace GGML_FA_ALL_QUANTS with GGML_FA_QUANTS, more control over what is compiled (#28079)

2026-09-09 12:50:08 +02:00

include

qwen4exp: add hc ops (#28901)

2026-09-16 16:00:01 +08:00

src

hexagon: accept the zeroed rope probe in supports_op (#28995)

2026-09-16 08:38:44 -07:00

.gitignore

vulkan : cmake integration (#8119)

2024-07-13 18:12:39 +02:00

CMakeLists.txt

ggml : bump version to 0.24.0 (ggml/1627)

2026-09-14 16:45:33 +03:00