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
107 lines
3.5 KiB
Protocol Buffer
107 lines
3.5 KiB
Protocol Buffer
syntax = "proto3";
|
|
|
|
option go_package = "jd.com/jd-infer/xllm;xllm";
|
|
package xllm.proto;
|
|
|
|
import "common.proto";
|
|
import "multimodal.proto";
|
|
import "tensor.proto";
|
|
import "embedding_data.proto";
|
|
|
|
message EmbeddingRequest {
|
|
// ID of the model to use. You can use the ListModels endpoint to list available models.
|
|
string model = 1;
|
|
|
|
// Input text to embed, encoded as a string or array of tokens. To embed multiple inputs in a single request, pass an array of strings or array of token arrays.
|
|
// The input must not exceed the max input tokens for the model (8192 tokens for text-embedding-ada-002),
|
|
// cannot be an empty string, and any array must be 2048 dimensions or less. Example Python code for counting tokens.
|
|
// Some models may also impose a limit on total number of tokens summed across inputs.
|
|
string input = 2;
|
|
//oneof input {
|
|
// // string, The string that will be turned into an embedding.
|
|
// string input_str = 2;
|
|
//
|
|
// // array, The array of strings that will be turned into an embedding.
|
|
// repeated string input_arr_str = 3;
|
|
//
|
|
// // array, The array of integers that will be turned into an embedding.
|
|
// repeated int32 input_arr_int = 4;
|
|
//
|
|
// // array, The array of arrays containing integers that will be turned into an embedding.
|
|
// repeated repeated int32 input_arr_arr_int = 5;
|
|
//}
|
|
|
|
// The number of dimensions the resulting output embeddings should have. Only supported in text-embedding-3 and later models.
|
|
optional int32 dimensions = 6;
|
|
|
|
// The format to return the embeddings in. Can be either float or base64.
|
|
// [default = "float"]
|
|
optional string encoding_format = 7;
|
|
|
|
// A unique identifier representing your end-user, which can help OpenAI to monitor and detect abuse.
|
|
optional string user = 8;
|
|
|
|
optional string service_request_id = 9;
|
|
|
|
optional bool add_special_tokens = 10;
|
|
}
|
|
|
|
message EmbeddingResponseData {
|
|
// The index of the embedding in the input array.
|
|
int32 index = 1;
|
|
|
|
// The object type of the embedding.
|
|
// [default = "embedding"]
|
|
string object = 2;
|
|
|
|
// The embedding vector.
|
|
repeated float embedding = 3;
|
|
repeated Embedding mm_embeddings = 4;
|
|
//oneof embedding {
|
|
// // float, The embedding vector as an array of floats.
|
|
// repeated float float_values = 3;
|
|
//
|
|
// // string, The embedding vector as a base64 encoded string.
|
|
// string base64_data = 4;
|
|
//}
|
|
}
|
|
|
|
message EmbeddingResponse {
|
|
// The ID of the embedding response.
|
|
string id = 1;
|
|
|
|
// The object type of the embedding response.
|
|
// [default = "list"]
|
|
string object = 2;
|
|
|
|
// The Unix timestamp of when the embedding response was created.
|
|
int64 created = 3;
|
|
|
|
// The model used to generate the embedding response.
|
|
string model = 4;
|
|
|
|
// The list of embedding response data.
|
|
repeated EmbeddingResponseData data = 5;
|
|
|
|
// Usage information for the embedding response.
|
|
Usage usage = 6;
|
|
}
|
|
|
|
message MMEmbeddingRequest {
|
|
// ID of the model to use. You can use the ListModels endpoint to list available models.
|
|
string model = 1;
|
|
|
|
repeated MMChatMessage messages = 3;
|
|
|
|
// The number of dimensions the resulting output embeddings should have. Only supported in text-embedding-3 and later models.
|
|
optional int32 dimensions = 6;
|
|
|
|
// The format to return the embeddings in. Can be either float or base64.
|
|
// [default = "float"]
|
|
optional string encoding_format = 7;
|
|
|
|
// A unique identifier representing your end-user, which can help OpenAI to monitor and detect abuse.
|
|
optional string user = 8;
|
|
|
|
optional string service_request_id = 9;
|
|
} |