vbatts/llama.cpp
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

Formatting changes

Browse source
This commit is contained in:
Phillip Kravtsov 2023-10-02 10:26:39 -07:00
parent e6bf87f785
commit cd4d3df820
3 changed files with 9 additions and 11 deletions
Download patch file Download diff file Expand all files Collapse all files

3
ggml-metal.m
View file

@ -870,8 +870,8 @@ void ggml_metal_graph_compute(
} break; } break;
case GGML_OP_MUL: case GGML_OP_MUL:
{ {
GGML_ASSERT(ggml_is_contiguous(src1));
GGML_ASSERT(ggml_is_contiguous(src0)); GGML_ASSERT(ggml_is_contiguous(src0));
GGML_ASSERT(ggml_is_contiguous(src1));
// utilize float4 // utilize float4
GGML_ASSERT(ne00 % 4 == 0); GGML_ASSERT(ne00 % 4 == 0);
@ -954,7 +954,6 @@ void ggml_metal_graph_compute(
[encoder setBuffer:id_src0 offset:offs_src0 atIndex:0]; [encoder setBuffer:id_src0 offset:offs_src0 atIndex:0];
[encoder setBuffer:id_dst offset:offs_dst atIndex:1]; [encoder setBuffer:id_dst offset:offs_dst atIndex:1];
//GGML_METAL_LOG_WARN("%s: node %3d, op = %8s dispatching \n", __func__, i, ggml_op_name(dst->op));
const int64_t n = ggml_nelements(dst); const int64_t n = ggml_nelements(dst);
[encoder dispatchThreadgroups:MTLSizeMake(n, 1, 1) threadsPerThreadgroup:MTLSizeMake(1, 1, 1)]; [encoder dispatchThreadgroups:MTLSizeMake(n, 1, 1) threadsPerThreadgroup:MTLSizeMake(1, 1, 1)];
} break; } break;

3
llama.cpp
View file

@ -4017,8 +4017,7 @@ static struct ggml_cgraph * llm_build_starcoder(
static struct ggml_cgraph * llm_build_persimmon( static struct ggml_cgraph * llm_build_persimmon(
llama_context & lctx, llama_context & lctx,
const llama_batch & batch const llama_batch & batch) {
) {
const auto & model = lctx.model; const auto & model = lctx.model;
const auto & hparams = model.hparams; const auto & hparams = model.hparams;