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

backends : add dev messages to support rope freq. factors

Browse source
This commit is contained in:
Georgi Gerganov 2024-05-16 13:23:30 +03:00
parent 471d8170bc
commit 352c3859a7
No known key found for this signature in database
GPG key ID: 449E073F9DC10735
3 changed files with 11 additions and 0 deletions
Download patch file Download diff file Expand all files Collapse all files

4
ggml-kompute.cpp
View file

@ -1677,6 +1677,10 @@ static void ggml_vk_graph_compute(struct ggml_kompute_context * ctx, struct ggml
} break; } break;
case GGML_OP_ROPE: case GGML_OP_ROPE:
{ {
#pragma message("TODO: implement phi3 frequency factors support");
#pragma message(" https://github.com/ggerganov/llama.cpp/pull/7225");
GGML_ASSERT(dst->src[2] == nullptr && "phi3 frequency factors not implemented yet");
GGML_ASSERT(ne10 == ne02); GGML_ASSERT(ne10 == ne02);
GGML_ASSERT(src0t == dstt); GGML_ASSERT(src0t == dstt);
// const int n_past = ((int32_t *) dst->op_params)[0]; // const int n_past = ((int32_t *) dst->op_params)[0];

3
ggml-sycl.cpp
View file

@ -14454,6 +14454,9 @@ inline void ggml_sycl_op_rope(const ggml_tensor *src0, const ggml_tensor *src1,
ggml_tensor *dst, const float *src0_dd, ggml_tensor *dst, const float *src0_dd,
const float *src1_dd, float *dst_dd, const float *src1_dd, float *dst_dd,
const dpct::queue_ptr &main_stream) { const dpct::queue_ptr &main_stream) {
#pragma message("TODO: implement phi3 frequency factors support");
#pragma message(" https://github.com/ggerganov/llama.cpp/pull/7225");
GGML_ASSERT(dst->src[2] == nullptr && "phi3 frequency factors not implemented yet");
GGML_ASSERT(src0->type == GGML_TYPE_F32 || src0->type == GGML_TYPE_F16); GGML_ASSERT(src0->type == GGML_TYPE_F32 || src0->type == GGML_TYPE_F16);
GGML_ASSERT( dst->type == GGML_TYPE_F32 || dst->type == GGML_TYPE_F16); GGML_ASSERT( dst->type == GGML_TYPE_F32 || dst->type == GGML_TYPE_F16);

4
ggml-vulkan.cpp
View file

@ -4238,6 +4238,10 @@ static void ggml_vk_soft_max(ggml_backend_vk_context * ctx, vk_context * subctx,
} }
static void ggml_vk_rope(ggml_backend_vk_context * ctx, vk_context * subctx, const ggml_tensor * src0, const ggml_tensor * src1, ggml_tensor * dst) { static void ggml_vk_rope(ggml_backend_vk_context * ctx, vk_context * subctx, const ggml_tensor * src0, const ggml_tensor * src1, ggml_tensor * dst) {
#pragma message("TODO: implement phi3 frequency factors support");
#pragma message(" https://github.com/ggerganov/llama.cpp/pull/7225");
GGML_ASSERT(dst->src[2] == nullptr && "phi3 frequency factors not implemented yet");
const int n_dims = ((int32_t *) dst->op_params)[1]; const int n_dims = ((int32_t *) dst->op_params)[1];
const int mode = ((int32_t *) dst->op_params)[2]; const int mode = ((int32_t *) dst->op_params)[2];
// const int n_ctx = ((int32_t *) dst->op_params)[3]; // const int n_ctx = ((int32_t *) dst->op_params)[3];