happyz
/
llama.cpp
mirror of https://github.com/ggerganov/llama.cpp.git
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity

fix: fix assertion

This commit is contained in:
hongruichen 2024-11-29 23:37:52 +08:00
parent 09efaa389e
commit c5e6549331
2 changed files with 3 additions and 1 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
ggml/src/ggml-qnn/backend-ops.cpp
View File

@ -666,6 +666,8 @@ bool ggml_qnn_supports_op(ggml_backend_qnn_device_context *ctx, const ggml_tenso
case GGML_OP_MUL_MAT: case GGML_OP_MUL_MAT:
return ggml_qnn_supports_matmul_op(ctx, op); return ggml_qnn_supports_matmul_op(ctx, op);
case GGML_OP_VIEW:
return true;
default: default:
return false; return false;
} }

2
ggml/src/ggml-qnn/ggml-qnn.cpp
View File

@ -227,7 +227,7 @@ size_t ggml_backend_qnn_buffer_type_get_alignment(ggml_backend_buffer_type_t buf
size_t ggml_backend_qnn_buffer_type_get_max_size(ggml_backend_buffer_type_t buft) { size_t ggml_backend_qnn_buffer_type_get_max_size(ggml_backend_buffer_type_t buft) {
GGML_UNUSED(buft); GGML_UNUSED(buft);
// TODO: get the max size from device // TODO: get the max size from device
return (1024 * 1024 * 1024); return (2 * 1024 * 1024 * 1024);
} }
bool ggml_backend_qnn_buffer_is_host(ggml_backend_buffer_type_t buft) { bool ggml_backend_qnn_buffer_is_host(ggml_backend_buffer_type_t buft) {