002f9879b2
Replaces cherry-picked upstream_ref with complete source trees. xllm/ — Iluvatar official C++ inference engine (15MB, 1470 files) Complete: kernels → layers → models → runtime → scheduler → api Excluded: .git, binary images, third_party submodule checkouts ds_vllm/ — Iluvatar official vllm fork (8MB, 703 files) Included: csrc/ (ALL CUDA kernels), fused_moe/, qwen3_5 model, _custom_ops Excluded: tests, benchmarks, docs, examples (not needed for reference) Critical call chains now fully traceable: MoE: moe_topk_softmax_kernels.cuh → ixformer.h → fused_moe.cpp → layer GDN: qwen3_gated_delta_net_base.cpp → qwen3_5_gated_delta_net.cpp Attention: ixformer.h → xllm_paged_attention → attention.cpp
35 lines
891 B
Protocol Buffer
35 lines
891 B
Protocol Buffer
syntax = "proto3";
|
|
|
|
option go_package = "jd.com/jd-infer/xllm;xllm";
|
|
package xllm.proto;
|
|
|
|
message Parameter
|
|
{
|
|
optional bool bool_param = 1;
|
|
optional int64 int64_param = 2;
|
|
optional string string_param = 3;
|
|
optional double double_param = 4;
|
|
optional uint64 uint64_param = 5;
|
|
}
|
|
|
|
message TensorContents
|
|
{
|
|
repeated bool bool_contents = 1;
|
|
repeated int32 int_contents = 2;
|
|
repeated int64 int64_contents = 3;
|
|
repeated uint32 uint_contents = 4;
|
|
repeated uint64 uint64_contents = 5;
|
|
repeated float fp32_contents = 6;
|
|
repeated double fp64_contents = 7;
|
|
repeated bytes string_contents = 8;
|
|
bytes bytes_contents = 9;
|
|
}
|
|
|
|
// TODO: rec.proto contains tensor definitions; refactor to separate tensor proto.
|
|
message Tensor {
|
|
string name = 1;
|
|
string datatype = 2;
|
|
repeated int64 shape = 3;
|
|
TensorContents contents = 4;
|
|
map<string, Parameter> parameters = 5;
|
|
} |