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
54 lines
1.3 KiB
Protocol Buffer
54 lines
1.3 KiB
Protocol Buffer
syntax = "proto3";
|
|
|
|
option go_package = "jd.com/jd-infer/xllm;xllm";
|
|
package xllm.proto;
|
|
|
|
import "common.proto";
|
|
import "completion.proto";
|
|
|
|
message SampleSelector {
|
|
// the selector type. V1 only supports "literal".
|
|
string type = 1;
|
|
|
|
// the literal value to match against the prompt.
|
|
string value = 2;
|
|
}
|
|
|
|
// Next ID: 6
|
|
message SampleRequest {
|
|
// ID of the model to use. You can use the ListModels endpoint to list available models.
|
|
string model = 1;
|
|
|
|
// the full prompt to sample from.
|
|
string prompt = 2;
|
|
|
|
// selector for locating sample positions in the prompt.
|
|
SampleSelector selector = 3;
|
|
|
|
// the number of log probabilities to include in the response, between [1, 5].
|
|
optional uint32 logprobs = 4;
|
|
|
|
// unique request identifier for traceability.
|
|
optional string request_id = 5;
|
|
}
|
|
|
|
message SampleResponse {
|
|
// unique id for the sample request.
|
|
string id = 1;
|
|
|
|
// the object type, which is always "sample_completion".
|
|
string object = 2;
|
|
|
|
// the unix timestamp (in seconds) of when the sample response was created.
|
|
uint32 created = 3;
|
|
|
|
// the model used for the sample response.
|
|
string model = 4;
|
|
|
|
// list of sample choices for selector matches in prompt order.
|
|
repeated Choice choices = 5;
|
|
|
|
// usage statistics for the sample request.
|
|
Usage usage = 6;
|
|
}
|