diff options
Diffstat (limited to 'ggml/src/ggml-cuda.cu')
-rw-r--r-- | ggml/src/ggml-cuda.cu | 5 |
1 files changed, 5 insertions, 0 deletions
diff --git a/ggml/src/ggml-cuda.cu b/ggml/src/ggml-cuda.cu index f594cd26..73ab0b73 100644 --- a/ggml/src/ggml-cuda.cu +++ b/ggml/src/ggml-cuda.cu @@ -24,6 +24,7 @@ #include "ggml-cuda/quantize.cuh" #include "ggml-cuda/rope.cuh" #include "ggml-cuda/scale.cuh" +#include "ggml-cuda/softcap.cuh" #include "ggml-cuda/softmax.cuh" #include "ggml-cuda/sumrows.cuh" #include "ggml-cuda/tsembd.cuh" @@ -2261,6 +2262,9 @@ static bool ggml_cuda_compute_forward(ggml_backend_cuda_context & ctx, struct gg case GGML_OP_SCALE: ggml_cuda_op_scale(ctx, dst); break; + case GGML_OP_SOFTCAP: + ggml_cuda_op_softcap(ctx, dst); + break; case GGML_OP_SQR: ggml_cuda_op_sqr(ctx, dst); break; @@ -2865,6 +2869,7 @@ GGML_CALL static bool ggml_backend_cuda_supports_op(ggml_backend_t backend, cons case GGML_OP_DIV: case GGML_OP_RMS_NORM: case GGML_OP_SCALE: + case GGML_OP_SOFTCAP: case GGML_OP_SQR: case GGML_OP_SQRT: case GGML_OP_CLAMP: |