mirror of
https://github.com/RYDE-WORK/llama.cpp.git
synced 2026-01-19 13:13:38 +08:00
* opencl: fix `ROPE` * opencl: fix `SOFT_MAX` * Add fp16 variant * opencl: enforce subgroup size for `soft_max`
* opencl: fix `ROPE` * opencl: fix `SOFT_MAX` * Add fp16 variant * opencl: enforce subgroup size for `soft_max`