char-fork/llama.cpp
git clone https://git.t4t.associates/char-fork/llama.cpp
3466812d1
char/aliceai-t5-moe
1#include "common.cuh" 2 3void ggml_cuda_op_moe_weighted_reduction(ggml_backend_cuda_context & ctx, 4 const ggml_tensor * experts, 5 const ggml_tensor * expert_scale, 6 const ggml_tensor * weights, 7 ggml_tensor * dst);