danyarm commited on
Commit
a7ba2c7
·
verified ·
1 Parent(s): 1a9f950

Upload queries/queries_nvidia-tensorrt-triton.jsonl with huggingface_hub

Browse files
queries/queries_nvidia-tensorrt-triton.jsonl ADDED
@@ -0,0 +1,167 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {"id": "nvidia-tensorrt-triton_T1_41", "domain": "nvidia-tensorrt-triton", "type": "T1_entity", "query": "What is Trt Llm Backend?", "ground_truth": ["Trt Llm Backend", "PLATFORM"], "concept_id": 41, "hop_depth": 0}
2
+ {"id": "nvidia-tensorrt-triton_T1_8", "domain": "nvidia-tensorrt-triton", "type": "T1_entity", "query": "What is Precision Modes?", "ground_truth": ["Precision Modes", "CONCEPT"], "concept_id": 8, "hop_depth": 0}
3
+ {"id": "nvidia-tensorrt-triton_T1_2", "domain": "nvidia-tensorrt-triton", "type": "T1_entity", "query": "What is Builder System?", "ground_truth": ["Builder System", "TOOL"], "concept_id": 2, "hop_depth": 0}
4
+ {"id": "nvidia-tensorrt-triton_T1_48", "domain": "nvidia-tensorrt-triton", "type": "T1_entity", "query": "What is Response Caching?", "ground_truth": ["Response Caching", "CONCEPT"], "concept_id": 48, "hop_depth": 0}
5
+ {"id": "nvidia-tensorrt-triton_T1_18", "domain": "nvidia-tensorrt-triton", "type": "T1_entity", "query": "What is Qat?", "ground_truth": ["Qat", "WORKFLOW"], "concept_id": 18, "hop_depth": 0}
6
+ {"id": "nvidia-tensorrt-triton_T1_16", "domain": "nvidia-tensorrt-triton", "type": "T1_entity", "query": "What is Packed Nhd Layout?", "ground_truth": ["Packed Nhd Layout", "STANDARD"], "concept_id": 16, "hop_depth": 0}
7
+ {"id": "nvidia-tensorrt-triton_T1_15", "domain": "nvidia-tensorrt-triton", "type": "T1_entity", "query": "What is Dynamic Shapes?", "ground_truth": ["Dynamic Shapes", "CONCEPT"], "concept_id": 15, "hop_depth": 0}
8
+ {"id": "nvidia-tensorrt-triton_T1_9", "domain": "nvidia-tensorrt-triton", "type": "T1_entity", "query": "What is Layer Fusion?", "ground_truth": ["Layer Fusion", "ALGORITH"], "concept_id": 9, "hop_depth": 0}
9
+ {"id": "nvidia-tensorrt-triton_T1_7", "domain": "nvidia-tensorrt-triton", "type": "T1_entity", "query": "What is Ipluginv3?", "ground_truth": ["Ipluginv3", "API"], "concept_id": 7, "hop_depth": 0}
10
+ {"id": "nvidia-tensorrt-triton_T1_35", "domain": "nvidia-tensorrt-triton", "type": "T1_entity", "query": "What is Triton C Api?", "ground_truth": ["Triton C Api", "API"], "concept_id": 35, "hop_depth": 0}
11
+ {"id": "nvidia-tensorrt-triton_T1_6", "domain": "nvidia-tensorrt-triton", "type": "T1_entity", "query": "What is Strong Typing Model?", "ground_truth": ["Strong Typing Model", "STANDARD"], "concept_id": 6, "hop_depth": 0}
12
+ {"id": "nvidia-tensorrt-triton_T1_38", "domain": "nvidia-tensorrt-triton", "type": "T1_entity", "query": "What is Sequence Batching?", "ground_truth": ["Sequence Batching", "ALGORITH"], "concept_id": 38, "hop_depth": 0}
13
+ {"id": "nvidia-tensorrt-triton_T1_28", "domain": "nvidia-tensorrt-triton", "type": "T1_entity", "query": "What is Onnx Import?", "ground_truth": ["Onnx Import", "WORKFLOW"], "concept_id": 28, "hop_depth": 0}
14
+ {"id": "nvidia-tensorrt-triton_T1_3", "domain": "nvidia-tensorrt-triton", "type": "T1_entity", "query": "What is Runtime Engine?", "ground_truth": ["Runtime Engine", "PLATFORM"], "concept_id": 3, "hop_depth": 0}
15
+ {"id": "nvidia-tensorrt-triton_T1_49", "domain": "nvidia-tensorrt-triton", "type": "T1_entity", "query": "What is Multi Lora?", "ground_truth": ["Multi Lora", "WORKFLOW"], "concept_id": 49, "hop_depth": 0}
16
+ {"id": "nvidia-tensorrt-triton_T1_51", "domain": "nvidia-tensorrt-triton", "type": "T1_entity", "query": "What is Kubernetes Deployment?", "ground_truth": ["Kubernetes Deployment", "WORKFLOW"], "concept_id": 51, "hop_depth": 0}
17
+ {"id": "nvidia-tensorrt-triton_T1_14", "domain": "nvidia-tensorrt-triton", "type": "T1_entity", "query": "What is Weight Streaming?", "ground_truth": ["Weight Streaming", "ALGORITH"], "concept_id": 14, "hop_depth": 0}
18
+ {"id": "nvidia-tensorrt-triton_T1_45", "domain": "nvidia-tensorrt-triton", "type": "T1_entity", "query": "What is Performance Analyzer?", "ground_truth": ["Performance Analyzer", "TOOL"], "concept_id": 45, "hop_depth": 0}
19
+ {"id": "nvidia-tensorrt-triton_T1_33", "domain": "nvidia-tensorrt-triton", "type": "T1_entity", "query": "What is Http Rest Api?", "ground_truth": ["Http Rest Api", "API"], "concept_id": 33, "hop_depth": 0}
20
+ {"id": "nvidia-tensorrt-triton_T1_37", "domain": "nvidia-tensorrt-triton", "type": "T1_entity", "query": "What is Dynamic Batching?", "ground_truth": ["Dynamic Batching", "ALGORITH"], "concept_id": 37, "hop_depth": 0}
21
+ {"id": "nvidia-tensorrt-triton_T1_47", "domain": "nvidia-tensorrt-triton", "type": "T1_entity", "query": "What is Genai Perf Analyzer?", "ground_truth": ["Genai Perf Analyzer", "TOOL"], "concept_id": 47, "hop_depth": 0}
22
+ {"id": "nvidia-tensorrt-triton_T1_43", "domain": "nvidia-tensorrt-triton", "type": "T1_entity", "query": "What is Onnxrt Backend?", "ground_truth": ["Onnxrt Backend", "PLATFORM"], "concept_id": 43, "hop_depth": 0}
23
+ {"id": "nvidia-tensorrt-triton_T1_23", "domain": "nvidia-tensorrt-triton", "type": "T1_entity", "query": "What is Fused Attention?", "ground_truth": ["Fused Attention", "ALGORITH"], "concept_id": 23, "hop_depth": 0}
24
+ {"id": "nvidia-tensorrt-triton_T1_21", "domain": "nvidia-tensorrt-triton", "type": "T1_entity", "query": "What is Moe Blackwell?", "ground_truth": ["Moe Blackwell", "ALGORITH"], "concept_id": 21, "hop_depth": 0}
25
+ {"id": "nvidia-tensorrt-triton_T1_29", "domain": "nvidia-tensorrt-triton", "type": "T1_entity", "query": "What is Nvfp4 Gemm Fusion?", "ground_truth": ["Nvfp4 Gemm Fusion", "ALGORITH"], "concept_id": 29, "hop_depth": 0}
26
+ {"id": "nvidia-tensorrt-triton_T1_31", "domain": "nvidia-tensorrt-triton", "type": "T1_entity", "query": "What is Triton Server?", "ground_truth": ["Triton Server", "PLATFORM"], "concept_id": 31, "hop_depth": 0}
27
+ {"id": "nvidia-tensorrt-triton_T1_42", "domain": "nvidia-tensorrt-triton", "type": "T1_entity", "query": "What is Vllm Backend?", "ground_truth": ["Vllm Backend", "PLATFORM"], "concept_id": 42, "hop_depth": 0}
28
+ {"id": "nvidia-tensorrt-triton_T1_50", "domain": "nvidia-tensorrt-triton", "type": "T1_entity", "query": "What is Speculative Decoding?", "ground_truth": ["Speculative Decoding", "ALGORITH"], "concept_id": 50, "hop_depth": 0}
29
+ {"id": "nvidia-tensorrt-triton_T1_34", "domain": "nvidia-tensorrt-triton", "type": "T1_entity", "query": "What is Grpc Api?", "ground_truth": ["Grpc Api", "API"], "concept_id": 34, "hop_depth": 0}
30
+ {"id": "nvidia-tensorrt-triton_T1_19", "domain": "nvidia-tensorrt-triton", "type": "T1_entity", "query": "What is Explicit Quantization?", "ground_truth": ["Explicit Quantization", "WORKFLOW"], "concept_id": 19, "hop_depth": 0}
31
+ {"id": "nvidia-tensorrt-triton_T1_44", "domain": "nvidia-tensorrt-triton", "type": "T1_entity", "query": "What is Kserve Spec?", "ground_truth": ["Kserve Spec", "STANDARD"], "concept_id": 44, "hop_depth": 0}
32
+ {"id": "nvidia-tensorrt-triton_T1_1", "domain": "nvidia-tensorrt-triton", "type": "T1_entity", "query": "What is Tensorrt Sdk?", "ground_truth": ["Tensorrt Sdk", "SDK"], "concept_id": 1, "hop_depth": 0}
33
+ {"id": "nvidia-tensorrt-triton_T1_36", "domain": "nvidia-tensorrt-triton", "type": "T1_entity", "query": "What is Backend C Api?", "ground_truth": ["Backend C Api", "API"], "concept_id": 36, "hop_depth": 0}
34
+ {"id": "nvidia-tensorrt-triton_T1_25", "domain": "nvidia-tensorrt-triton", "type": "T1_entity", "query": "What is Multi Device Inference?", "ground_truth": ["Multi Device Inference", "WORKFLOW"], "concept_id": 25, "hop_depth": 0}
35
+ {"id": "nvidia-tensorrt-triton_T1_11", "domain": "nvidia-tensorrt-triton", "type": "T1_entity", "query": "What is Timing Cache?", "ground_truth": ["Timing Cache", "CONCEPT"], "concept_id": 11, "hop_depth": 0}
36
+ {"id": "nvidia-tensorrt-triton_T1_39", "domain": "nvidia-tensorrt-triton", "type": "T1_entity", "query": "What is Ensemble Model?", "ground_truth": ["Ensemble Model", "CONCEPT"], "concept_id": 39, "hop_depth": 0}
37
+ {"id": "nvidia-tensorrt-triton_T1_40", "domain": "nvidia-tensorrt-triton", "type": "T1_entity", "query": "What is Bls?", "ground_truth": ["Bls", "TOOL"], "concept_id": 40, "hop_depth": 0}
38
+ {"id": "nvidia-tensorrt-triton_T1_4", "domain": "nvidia-tensorrt-triton", "type": "T1_entity", "query": "What is Trt Cpp Api?", "ground_truth": ["Trt Cpp Api", "API"], "concept_id": 4, "hop_depth": 0}
39
+ {"id": "nvidia-tensorrt-triton_T1_13", "domain": "nvidia-tensorrt-triton", "type": "T1_entity", "query": "What is Engine Refitting?", "ground_truth": ["Engine Refitting", "WORKFLOW"], "concept_id": 13, "hop_depth": 0}
40
+ {"id": "nvidia-tensorrt-triton_T1_22", "domain": "nvidia-tensorrt-triton", "type": "T1_entity", "query": "What is Rotary Pos Embed?", "ground_truth": ["Rotary Pos Embed", "ALGORITH"], "concept_id": 22, "hop_depth": 0}
41
+ {"id": "nvidia-tensorrt-triton_T1_32", "domain": "nvidia-tensorrt-triton", "type": "T1_entity", "query": "What is Model Repository?", "ground_truth": ["Model Repository", "CONCEPT"], "concept_id": 32, "hop_depth": 0}
42
+ {"id": "nvidia-tensorrt-triton_T1_17", "domain": "nvidia-tensorrt-triton", "type": "T1_entity", "query": "What is Ptq?", "ground_truth": ["Ptq", "WORKFLOW"], "concept_id": 17, "hop_depth": 0}
43
+ {"id": "nvidia-tensorrt-triton_T1_30", "domain": "nvidia-tensorrt-triton", "type": "T1_entity", "query": "What is Api Capture Replay?", "ground_truth": ["Api Capture Replay", "TOOL"], "concept_id": 30, "hop_depth": 0}
44
+ {"id": "nvidia-tensorrt-triton_T1_10", "domain": "nvidia-tensorrt-triton", "type": "T1_entity", "query": "What is Global Perf Tuner?", "ground_truth": ["Global Perf Tuner", "TOOL"], "concept_id": 10, "hop_depth": 0}
45
+ {"id": "nvidia-tensorrt-triton_T1_27", "domain": "nvidia-tensorrt-triton", "type": "T1_entity", "query": "What is Dla?", "ground_truth": ["Dla", "PLATFORM"], "concept_id": 27, "hop_depth": 0}
46
+ {"id": "nvidia-tensorrt-triton_T1_46", "domain": "nvidia-tensorrt-triton", "type": "T1_entity", "query": "What is Model Analyzer?", "ground_truth": ["Model Analyzer", "TOOL"], "concept_id": 46, "hop_depth": 0}
47
+ {"id": "nvidia-tensorrt-triton_T1_5", "domain": "nvidia-tensorrt-triton", "type": "T1_entity", "query": "What is Trt Python Api?", "ground_truth": ["Trt Python Api", "API"], "concept_id": 5, "hop_depth": 0}
48
+ {"id": "nvidia-tensorrt-triton_T1_12", "domain": "nvidia-tensorrt-triton", "type": "T1_entity", "query": "What is Cuda Graph?", "ground_truth": ["Cuda Graph", "ALGORITH"], "concept_id": 12, "hop_depth": 0}
49
+ {"id": "nvidia-tensorrt-triton_T1_20", "domain": "nvidia-tensorrt-triton", "type": "T1_entity", "query": "What is Kv Cache?", "ground_truth": ["Kv Cache", "ALGORITH"], "concept_id": 20, "hop_depth": 0}
50
+ {"id": "nvidia-tensorrt-triton_T1_24", "domain": "nvidia-tensorrt-triton", "type": "T1_entity", "query": "What is Iattention Layer?", "ground_truth": ["Iattention Layer", "API"], "concept_id": 24, "hop_depth": 0}
51
+ {"id": "nvidia-tensorrt-triton_T2_39", "domain": "nvidia-tensorrt-triton", "type": "T2_dependency", "query": "What are the prerequisites for Ensemble Model?", "ground_truth": ["Triton Server"], "concept_id": 39, "hop_depth": 1}
52
+ {"id": "nvidia-tensorrt-triton_T2_9", "domain": "nvidia-tensorrt-triton", "type": "T2_dependency", "query": "What are the prerequisites for Layer Fusion?", "ground_truth": ["Builder System"], "concept_id": 9, "hop_depth": 1}
53
+ {"id": "nvidia-tensorrt-triton_T2_29", "domain": "nvidia-tensorrt-triton", "type": "T2_dependency", "query": "What are the prerequisites for Nvfp4 Gemm Fusion?", "ground_truth": ["Tensorrt Sdk", "Precision Modes"], "concept_id": 29, "hop_depth": 1}
54
+ {"id": "nvidia-tensorrt-triton_T2_7", "domain": "nvidia-tensorrt-triton", "type": "T2_dependency", "query": "What are the prerequisites for Ipluginv3?", "ground_truth": ["Tensorrt Sdk"], "concept_id": 7, "hop_depth": 1}
55
+ {"id": "nvidia-tensorrt-triton_T2_40", "domain": "nvidia-tensorrt-triton", "type": "T2_dependency", "query": "What are the prerequisites for Bls?", "ground_truth": ["Ensemble Model"], "concept_id": 40, "hop_depth": 1}
56
+ {"id": "nvidia-tensorrt-triton_T2_20", "domain": "nvidia-tensorrt-triton", "type": "T2_dependency", "query": "What are the prerequisites for Kv Cache?", "ground_truth": ["Fused Attention"], "concept_id": 20, "hop_depth": 1}
57
+ {"id": "nvidia-tensorrt-triton_T2_45", "domain": "nvidia-tensorrt-triton", "type": "T2_dependency", "query": "What are the prerequisites for Performance Analyzer?", "ground_truth": ["Triton Server"], "concept_id": 45, "hop_depth": 1}
58
+ {"id": "nvidia-tensorrt-triton_T2_27", "domain": "nvidia-tensorrt-triton", "type": "T2_dependency", "query": "What are the prerequisites for Dla?", "ground_truth": ["Tensorrt Sdk"], "concept_id": 27, "hop_depth": 1}
59
+ {"id": "nvidia-tensorrt-triton_T2_41", "domain": "nvidia-tensorrt-triton", "type": "T2_dependency", "query": "What are the prerequisites for Trt Llm Backend?", "ground_truth": ["Tensorrt Sdk", "Backend C Api", "Kv Cache"], "concept_id": 41, "hop_depth": 1}
60
+ {"id": "nvidia-tensorrt-triton_T2_14", "domain": "nvidia-tensorrt-triton", "type": "T2_dependency", "query": "What are the prerequisites for Weight Streaming?", "ground_truth": ["Runtime Engine"], "concept_id": 14, "hop_depth": 1}
61
+ {"id": "nvidia-tensorrt-triton_T2_5", "domain": "nvidia-tensorrt-triton", "type": "T2_dependency", "query": "What are the prerequisites for Trt Python Api?", "ground_truth": ["Tensorrt Sdk", "Runtime Engine", "Strong Typing Model", "Builder System"], "concept_id": 5, "hop_depth": 1}
62
+ {"id": "nvidia-tensorrt-triton_T2_3", "domain": "nvidia-tensorrt-triton", "type": "T2_dependency", "query": "What are the prerequisites for Runtime Engine?", "ground_truth": ["Builder System"], "concept_id": 3, "hop_depth": 1}
63
+ {"id": "nvidia-tensorrt-triton_T2_16", "domain": "nvidia-tensorrt-triton", "type": "T2_dependency", "query": "What are the prerequisites for Packed Nhd Layout?", "ground_truth": ["Dynamic Shapes"], "concept_id": 16, "hop_depth": 1}
64
+ {"id": "nvidia-tensorrt-triton_T2_46", "domain": "nvidia-tensorrt-triton", "type": "T2_dependency", "query": "What are the prerequisites for Model Analyzer?", "ground_truth": ["Triton Server"], "concept_id": 46, "hop_depth": 1}
65
+ {"id": "nvidia-tensorrt-triton_T2_48", "domain": "nvidia-tensorrt-triton", "type": "T2_dependency", "query": "What are the prerequisites for Response Caching?", "ground_truth": ["Triton Server"], "concept_id": 48, "hop_depth": 1}
66
+ {"id": "nvidia-tensorrt-triton_T2_32", "domain": "nvidia-tensorrt-triton", "type": "T2_dependency", "query": "What are the prerequisites for Model Repository?", "ground_truth": ["Triton Server"], "concept_id": 32, "hop_depth": 1}
67
+ {"id": "nvidia-tensorrt-triton_T2_50", "domain": "nvidia-tensorrt-triton", "type": "T2_dependency", "query": "What are the prerequisites for Speculative Decoding?", "ground_truth": ["Vllm Backend"], "concept_id": 50, "hop_depth": 1}
68
+ {"id": "nvidia-tensorrt-triton_T2_35", "domain": "nvidia-tensorrt-triton", "type": "T2_dependency", "query": "What are the prerequisites for Triton C Api?", "ground_truth": ["Triton Server"], "concept_id": 35, "hop_depth": 1}
69
+ {"id": "nvidia-tensorrt-triton_T2_4", "domain": "nvidia-tensorrt-triton", "type": "T2_dependency", "query": "What are the prerequisites for Trt Cpp Api?", "ground_truth": ["Tensorrt Sdk", "Runtime Engine", "Strong Typing Model", "Builder System"], "concept_id": 4, "hop_depth": 1}
70
+ {"id": "nvidia-tensorrt-triton_T2_42", "domain": "nvidia-tensorrt-triton", "type": "T2_dependency", "query": "What are the prerequisites for Vllm Backend?", "ground_truth": ["Backend C Api"], "concept_id": 42, "hop_depth": 1}
71
+ {"id": "nvidia-tensorrt-triton_T2_10", "domain": "nvidia-tensorrt-triton", "type": "T2_dependency", "query": "What are the prerequisites for Global Perf Tuner?", "ground_truth": ["Timing Cache", "Builder System"], "concept_id": 10, "hop_depth": 1}
72
+ {"id": "nvidia-tensorrt-triton_T2_38", "domain": "nvidia-tensorrt-triton", "type": "T2_dependency", "query": "What are the prerequisites for Sequence Batching?", "ground_truth": ["Triton Server"], "concept_id": 38, "hop_depth": 1}
73
+ {"id": "nvidia-tensorrt-triton_T2_23", "domain": "nvidia-tensorrt-triton", "type": "T2_dependency", "query": "What are the prerequisites for Fused Attention?", "ground_truth": ["Tensorrt Sdk", "Rotary Pos Embed", "Layer Fusion"], "concept_id": 23, "hop_depth": 1}
74
+ {"id": "nvidia-tensorrt-triton_T2_13", "domain": "nvidia-tensorrt-triton", "type": "T2_dependency", "query": "What are the prerequisites for Engine Refitting?", "ground_truth": ["Runtime Engine"], "concept_id": 13, "hop_depth": 1}
75
+ {"id": "nvidia-tensorrt-triton_T2_36", "domain": "nvidia-tensorrt-triton", "type": "T2_dependency", "query": "What are the prerequisites for Backend C Api?", "ground_truth": ["Triton Server"], "concept_id": 36, "hop_depth": 1}
76
+ {"id": "nvidia-tensorrt-triton_T2_25", "domain": "nvidia-tensorrt-triton", "type": "T2_dependency", "query": "What are the prerequisites for Multi Device Inference?", "ground_truth": ["Runtime Engine", "Nccl Backend"], "concept_id": 25, "hop_depth": 1}
77
+ {"id": "nvidia-tensorrt-triton_T2_43", "domain": "nvidia-tensorrt-triton", "type": "T2_dependency", "query": "What are the prerequisites for Onnxrt Backend?", "ground_truth": ["Backend C Api", "Onnx Import"], "concept_id": 43, "hop_depth": 1}
78
+ {"id": "nvidia-tensorrt-triton_T2_8", "domain": "nvidia-tensorrt-triton", "type": "T2_dependency", "query": "What are the prerequisites for Precision Modes?", "ground_truth": ["Strong Typing Model"], "concept_id": 8, "hop_depth": 1}
79
+ {"id": "nvidia-tensorrt-triton_T2_30", "domain": "nvidia-tensorrt-triton", "type": "T2_dependency", "query": "What are the prerequisites for Api Capture Replay?", "ground_truth": ["Runtime Engine"], "concept_id": 30, "hop_depth": 1}
80
+ {"id": "nvidia-tensorrt-triton_T2_51", "domain": "nvidia-tensorrt-triton", "type": "T2_dependency", "query": "What are the prerequisites for Kubernetes Deployment?", "ground_truth": ["Triton Server"], "concept_id": 51, "hop_depth": 1}
81
+ {"id": "nvidia-tensorrt-triton_T2_24", "domain": "nvidia-tensorrt-triton", "type": "T2_dependency", "query": "What are the prerequisites for Iattention Layer?", "ground_truth": ["Kv Cache", "Fused Attention", "Packed Nhd Layout"], "concept_id": 24, "hop_depth": 1}
82
+ {"id": "nvidia-tensorrt-triton_T2_19", "domain": "nvidia-tensorrt-triton", "type": "T2_dependency", "query": "What are the prerequisites for Explicit Quantization?", "ground_truth": ["Strong Typing Model", "Precision Modes"], "concept_id": 19, "hop_depth": 1}
83
+ {"id": "nvidia-tensorrt-triton_T2_21", "domain": "nvidia-tensorrt-triton", "type": "T2_dependency", "query": "What are the prerequisites for Moe Blackwell?", "ground_truth": ["Tensorrt Sdk", "Precision Modes"], "concept_id": 21, "hop_depth": 1}
84
+ {"id": "nvidia-tensorrt-triton_T2_33", "domain": "nvidia-tensorrt-triton", "type": "T2_dependency", "query": "What are the prerequisites for Http Rest Api?", "ground_truth": ["Triton Server", "Kserve Spec"], "concept_id": 33, "hop_depth": 1}
85
+ {"id": "nvidia-tensorrt-triton_T2_18", "domain": "nvidia-tensorrt-triton", "type": "T2_dependency", "query": "What are the prerequisites for Qat?", "ground_truth": ["Precision Modes"], "concept_id": 18, "hop_depth": 1}
86
+ {"id": "nvidia-tensorrt-triton_T2_34", "domain": "nvidia-tensorrt-triton", "type": "T2_dependency", "query": "What are the prerequisites for Grpc Api?", "ground_truth": ["Triton Server", "Kserve Spec"], "concept_id": 34, "hop_depth": 1}
87
+ {"id": "nvidia-tensorrt-triton_T2_37", "domain": "nvidia-tensorrt-triton", "type": "T2_dependency", "query": "What are the prerequisites for Dynamic Batching?", "ground_truth": ["Triton Server"], "concept_id": 37, "hop_depth": 1}
88
+ {"id": "nvidia-tensorrt-triton_T2_47", "domain": "nvidia-tensorrt-triton", "type": "T2_dependency", "query": "What are the prerequisites for Genai Perf Analyzer?", "ground_truth": ["Triton Server"], "concept_id": 47, "hop_depth": 1}
89
+ {"id": "nvidia-tensorrt-triton_T2_31", "domain": "nvidia-tensorrt-triton", "type": "T2_dependency", "query": "What are the prerequisites for Triton Server?", "ground_truth": ["Triton C Api", "Model Repository", "Http Rest Api", "Grpc Api"], "concept_id": 31, "hop_depth": 1}
90
+ {"id": "nvidia-tensorrt-triton_T2_17", "domain": "nvidia-tensorrt-triton", "type": "T2_dependency", "query": "What are the prerequisites for Ptq?", "ground_truth": ["Builder System", "Precision Modes"], "concept_id": 17, "hop_depth": 1}
91
+ {"id": "nvidia-tensorrt-triton_T2_15", "domain": "nvidia-tensorrt-triton", "type": "T2_dependency", "query": "What are the prerequisites for Dynamic Shapes?", "ground_truth": ["Builder System", "Packed Nhd Layout"], "concept_id": 15, "hop_depth": 1}
92
+ {"id": "nvidia-tensorrt-triton_T2_1", "domain": "nvidia-tensorrt-triton", "type": "T2_dependency", "query": "What are the prerequisites for Tensorrt Sdk?", "ground_truth": ["Runtime Engine", "Builder System"], "concept_id": 1, "hop_depth": 1}
93
+ {"id": "nvidia-tensorrt-triton_T2_2", "domain": "nvidia-tensorrt-triton", "type": "T2_dependency", "query": "What are the prerequisites for Builder System?", "ground_truth": ["Tensorrt Sdk", "Onnx Import"], "concept_id": 2, "hop_depth": 1}
94
+ {"id": "nvidia-tensorrt-triton_T2_49", "domain": "nvidia-tensorrt-triton", "type": "T2_dependency", "query": "What are the prerequisites for Multi Lora?", "ground_truth": ["Vllm Backend"], "concept_id": 49, "hop_depth": 1}
95
+ {"id": "nvidia-tensorrt-triton_T2_12", "domain": "nvidia-tensorrt-triton", "type": "T2_dependency", "query": "What are the prerequisites for Cuda Graph?", "ground_truth": ["Runtime Engine"], "concept_id": 12, "hop_depth": 1}
96
+ {"id": "nvidia-tensorrt-triton_T2_11", "domain": "nvidia-tensorrt-triton", "type": "T2_dependency", "query": "What are the prerequisites for Timing Cache?", "ground_truth": ["Builder System"], "concept_id": 11, "hop_depth": 1}
97
+ {"id": "nvidia-tensorrt-triton_T3_28_25", "domain": "nvidia-tensorrt-triton", "type": "T3_path", "query": "What is the prerequisite chain from Onnx Import to Multi Device Inference?", "ground_truth": ["Multi Device Inference", "Runtime Engine", "Builder System", "Onnx Import"], "concept_id": 25, "hop_depth": 3, "path_ids": [25, 3, 2, 28]}
98
+ {"id": "nvidia-tensorrt-triton_T3_28_12", "domain": "nvidia-tensorrt-triton", "type": "T3_path", "query": "What is the prerequisite chain from Onnx Import to Cuda Graph?", "ground_truth": ["Cuda Graph", "Runtime Engine", "Builder System", "Onnx Import"], "concept_id": 12, "hop_depth": 3, "path_ids": [12, 3, 2, 28]}
99
+ {"id": "nvidia-tensorrt-triton_T3_44_38", "domain": "nvidia-tensorrt-triton", "type": "T3_path", "query": "What is the prerequisite chain from Kserve Spec to Sequence Batching?", "ground_truth": ["Sequence Batching", "Triton Server", "Http Rest Api", "Kserve Spec"], "concept_id": 38, "hop_depth": 3, "path_ids": [38, 31, 33, 44]}
100
+ {"id": "nvidia-tensorrt-triton_T3_28_7", "domain": "nvidia-tensorrt-triton", "type": "T3_path", "query": "What is the prerequisite chain from Onnx Import to Ipluginv3?", "ground_truth": ["Ipluginv3", "Tensorrt Sdk", "Builder System", "Onnx Import"], "concept_id": 7, "hop_depth": 3, "path_ids": [7, 1, 2, 28]}
101
+ {"id": "nvidia-tensorrt-triton_T3_28_41", "domain": "nvidia-tensorrt-triton", "type": "T3_path", "query": "What is the prerequisite chain from Onnx Import to Trt Llm Backend?", "ground_truth": ["Trt Llm Backend", "Tensorrt Sdk", "Builder System", "Onnx Import"], "concept_id": 41, "hop_depth": 3, "path_ids": [41, 1, 2, 28]}
102
+ {"id": "nvidia-tensorrt-triton_T3_28_41", "domain": "nvidia-tensorrt-triton", "type": "T3_path", "query": "What is the prerequisite chain from Onnx Import to Trt Llm Backend?", "ground_truth": ["Trt Llm Backend", "Tensorrt Sdk", "Builder System", "Onnx Import"], "concept_id": 41, "hop_depth": 3, "path_ids": [41, 1, 2, 28]}
103
+ {"id": "nvidia-tensorrt-triton_T3_44_48", "domain": "nvidia-tensorrt-triton", "type": "T3_path", "query": "What is the prerequisite chain from Kserve Spec to Response Caching?", "ground_truth": ["Response Caching", "Triton Server", "Http Rest Api", "Kserve Spec"], "concept_id": 48, "hop_depth": 3, "path_ids": [48, 31, 33, 44]}
104
+ {"id": "nvidia-tensorrt-triton_T3_44_41", "domain": "nvidia-tensorrt-triton", "type": "T3_path", "query": "What is the prerequisite chain from Kserve Spec to Trt Llm Backend?", "ground_truth": ["Trt Llm Backend", "Backend C Api", "Triton Server", "Http Rest Api", "Kserve Spec"], "concept_id": 41, "hop_depth": 4, "path_ids": [41, 36, 31, 33, 44]}
105
+ {"id": "nvidia-tensorrt-triton_T3_28_4", "domain": "nvidia-tensorrt-triton", "type": "T3_path", "query": "What is the prerequisite chain from Onnx Import to Trt Cpp Api?", "ground_truth": ["Trt Cpp Api", "Builder System", "Onnx Import"], "concept_id": 4, "hop_depth": 2, "path_ids": [4, 2, 28]}
106
+ {"id": "nvidia-tensorrt-triton_T3_6_17", "domain": "nvidia-tensorrt-triton", "type": "T3_path", "query": "What is the prerequisite chain from Strong Typing Model to Ptq?", "ground_truth": ["Ptq", "Precision Modes", "Strong Typing Model"], "concept_id": 17, "hop_depth": 2, "path_ids": [17, 8, 6]}
107
+ {"id": "nvidia-tensorrt-triton_T3_44_48", "domain": "nvidia-tensorrt-triton", "type": "T3_path", "query": "What is the prerequisite chain from Kserve Spec to Response Caching?", "ground_truth": ["Response Caching", "Triton Server", "Http Rest Api", "Kserve Spec"], "concept_id": 48, "hop_depth": 3, "path_ids": [48, 31, 33, 44]}
108
+ {"id": "nvidia-tensorrt-triton_T3_44_51", "domain": "nvidia-tensorrt-triton", "type": "T3_path", "query": "What is the prerequisite chain from Kserve Spec to Kubernetes Deployment?", "ground_truth": ["Kubernetes Deployment", "Triton Server", "Http Rest Api", "Kserve Spec"], "concept_id": 51, "hop_depth": 3, "path_ids": [51, 31, 33, 44]}
109
+ {"id": "nvidia-tensorrt-triton_T3_26_25", "domain": "nvidia-tensorrt-triton", "type": "T3_path", "query": "What is the prerequisite chain from Nccl Backend to Multi Device Inference?", "ground_truth": ["Multi Device Inference", "Nccl Backend"], "concept_id": 25, "hop_depth": 1, "path_ids": [25, 26]}
110
+ {"id": "nvidia-tensorrt-triton_T3_6_17", "domain": "nvidia-tensorrt-triton", "type": "T3_path", "query": "What is the prerequisite chain from Strong Typing Model to Ptq?", "ground_truth": ["Ptq", "Precision Modes", "Strong Typing Model"], "concept_id": 17, "hop_depth": 2, "path_ids": [17, 8, 6]}
111
+ {"id": "nvidia-tensorrt-triton_T3_44_38", "domain": "nvidia-tensorrt-triton", "type": "T3_path", "query": "What is the prerequisite chain from Kserve Spec to Sequence Batching?", "ground_truth": ["Sequence Batching", "Triton Server", "Http Rest Api", "Kserve Spec"], "concept_id": 38, "hop_depth": 3, "path_ids": [38, 31, 33, 44]}
112
+ {"id": "nvidia-tensorrt-triton_T3_6_17", "domain": "nvidia-tensorrt-triton", "type": "T3_path", "query": "What is the prerequisite chain from Strong Typing Model to Ptq?", "ground_truth": ["Ptq", "Precision Modes", "Strong Typing Model"], "concept_id": 17, "hop_depth": 2, "path_ids": [17, 8, 6]}
113
+ {"id": "nvidia-tensorrt-triton_T3_6_5", "domain": "nvidia-tensorrt-triton", "type": "T3_path", "query": "What is the prerequisite chain from Strong Typing Model to Trt Python Api?", "ground_truth": ["Trt Python Api", "Strong Typing Model"], "concept_id": 5, "hop_depth": 1, "path_ids": [5, 6]}
114
+ {"id": "nvidia-tensorrt-triton_T3_28_14", "domain": "nvidia-tensorrt-triton", "type": "T3_path", "query": "What is the prerequisite chain from Onnx Import to Weight Streaming?", "ground_truth": ["Weight Streaming", "Runtime Engine", "Builder System", "Onnx Import"], "concept_id": 14, "hop_depth": 3, "path_ids": [14, 3, 2, 28]}
115
+ {"id": "nvidia-tensorrt-triton_T3_44_40", "domain": "nvidia-tensorrt-triton", "type": "T3_path", "query": "What is the prerequisite chain from Kserve Spec to Bls?", "ground_truth": ["Bls", "Ensemble Model", "Triton Server", "Http Rest Api", "Kserve Spec"], "concept_id": 40, "hop_depth": 4, "path_ids": [40, 39, 31, 33, 44]}
116
+ {"id": "nvidia-tensorrt-triton_T3_28_17", "domain": "nvidia-tensorrt-triton", "type": "T3_path", "query": "What is the prerequisite chain from Onnx Import to Ptq?", "ground_truth": ["Ptq", "Builder System", "Onnx Import"], "concept_id": 17, "hop_depth": 2, "path_ids": [17, 2, 28]}
117
+ {"id": "nvidia-tensorrt-triton_T3_28_14", "domain": "nvidia-tensorrt-triton", "type": "T3_path", "query": "What is the prerequisite chain from Onnx Import to Weight Streaming?", "ground_truth": ["Weight Streaming", "Runtime Engine", "Builder System", "Onnx Import"], "concept_id": 14, "hop_depth": 3, "path_ids": [14, 3, 2, 28]}
118
+ {"id": "nvidia-tensorrt-triton_T3_28_24", "domain": "nvidia-tensorrt-triton", "type": "T3_path", "query": "What is the prerequisite chain from Onnx Import to Iattention Layer?", "ground_truth": ["Iattention Layer", "Fused Attention", "Tensorrt Sdk", "Builder System", "Onnx Import"], "concept_id": 24, "hop_depth": 4, "path_ids": [24, 23, 1, 2, 28]}
119
+ {"id": "nvidia-tensorrt-triton_T3_28_27", "domain": "nvidia-tensorrt-triton", "type": "T3_path", "query": "What is the prerequisite chain from Onnx Import to Dla?", "ground_truth": ["Dla", "Tensorrt Sdk", "Builder System", "Onnx Import"], "concept_id": 27, "hop_depth": 3, "path_ids": [27, 1, 2, 28]}
120
+ {"id": "nvidia-tensorrt-triton_T3_6_5", "domain": "nvidia-tensorrt-triton", "type": "T3_path", "query": "What is the prerequisite chain from Strong Typing Model to Trt Python Api?", "ground_truth": ["Trt Python Api", "Strong Typing Model"], "concept_id": 5, "hop_depth": 1, "path_ids": [5, 6]}
121
+ {"id": "nvidia-tensorrt-triton_T3_6_17", "domain": "nvidia-tensorrt-triton", "type": "T3_path", "query": "What is the prerequisite chain from Strong Typing Model to Ptq?", "ground_truth": ["Ptq", "Precision Modes", "Strong Typing Model"], "concept_id": 17, "hop_depth": 2, "path_ids": [17, 8, 6]}
122
+ {"id": "nvidia-tensorrt-triton_T4_SDK", "domain": "nvidia-tensorrt-triton", "type": "T4_aggregate", "query": "List all SDK concepts in this knowledge graph", "ground_truth": ["Tensorrt Sdk"], "taxonomy_id": "SDK", "hop_depth": 0}
123
+ {"id": "nvidia-tensorrt-triton_T4_TOOL", "domain": "nvidia-tensorrt-triton", "type": "T4_aggregate", "query": "List all TOOL concepts in this knowledge graph", "ground_truth": ["Builder System", "Global Perf Tuner", "Api Capture Replay", "Bls", "Performance Analyzer", "Model Analyzer", "Genai Perf Analyzer"], "taxonomy_id": "TOOL", "hop_depth": 0}
124
+ {"id": "nvidia-tensorrt-triton_T4_PLATFORM", "domain": "nvidia-tensorrt-triton", "type": "T4_aggregate", "query": "List all PLATFORM concepts in this knowledge graph", "ground_truth": ["Runtime Engine", "Nccl Backend", "Dla", "Triton Server", "Trt Llm Backend", "Vllm Backend", "Onnxrt Backend"], "taxonomy_id": "PLATFORM", "hop_depth": 0}
125
+ {"id": "nvidia-tensorrt-triton_T4_API", "domain": "nvidia-tensorrt-triton", "type": "T4_aggregate", "query": "List all API concepts in this knowledge graph", "ground_truth": ["Trt Cpp Api", "Trt Python Api", "Ipluginv3", "Iattention Layer", "Http Rest Api", "Grpc Api", "Triton C Api", "Backend C Api"], "taxonomy_id": "API", "hop_depth": 0}
126
+ {"id": "nvidia-tensorrt-triton_T4_STANDARD", "domain": "nvidia-tensorrt-triton", "type": "T4_aggregate", "query": "List all STANDARD concepts in this knowledge graph", "ground_truth": ["Strong Typing Model", "Packed Nhd Layout", "Kserve Spec"], "taxonomy_id": "STANDARD", "hop_depth": 0}
127
+ {"id": "nvidia-tensorrt-triton_T4_CONCEPT", "domain": "nvidia-tensorrt-triton", "type": "T4_aggregate", "query": "List all CONCEPT concepts in this knowledge graph", "ground_truth": ["Precision Modes", "Timing Cache", "Dynamic Shapes", "Model Repository", "Ensemble Model", "Response Caching"], "taxonomy_id": "CONCEPT", "hop_depth": 0}
128
+ {"id": "nvidia-tensorrt-triton_T4_ALGORITH", "domain": "nvidia-tensorrt-triton", "type": "T4_aggregate", "query": "List all ALGORITH concepts in this knowledge graph", "ground_truth": ["Layer Fusion", "Cuda Graph", "Weight Streaming", "Kv Cache", "Moe Blackwell", "Rotary Pos Embed", "Fused Attention", "Nvfp4 Gemm Fusion", "Dynamic Batching", "Sequence Batching", "Speculative Decoding"], "taxonomy_id": "ALGORITH", "hop_depth": 0}
129
+ {"id": "nvidia-tensorrt-triton_T4_WORKFLOW", "domain": "nvidia-tensorrt-triton", "type": "T4_aggregate", "query": "List all WORKFLOW concepts in this knowledge graph", "ground_truth": ["Engine Refitting", "Ptq", "Qat", "Explicit Quantization", "Multi Device Inference", "Onnx Import", "Multi Lora", "Kubernetes Deployment"], "taxonomy_id": "WORKFLOW", "hop_depth": 0}
130
+ {"id": "nvidia-tensorrt-triton_T5_16_15", "domain": "nvidia-tensorrt-triton", "type": "T5_cross_concept", "query": "How does Packed Nhd Layout relate to Dynamic Shapes?", "ground_truth": ["Packed Nhd Layout", "Dynamic Shapes"], "concept_id_a": 16, "concept_id_b": 15, "hop_depth": 1}
131
+ {"id": "nvidia-tensorrt-triton_T5_51_31", "domain": "nvidia-tensorrt-triton", "type": "T5_cross_concept", "query": "How does Kubernetes Deployment relate to Triton Server?", "ground_truth": ["Kubernetes Deployment", "Triton Server"], "concept_id_a": 51, "concept_id_b": 31, "hop_depth": 1}
132
+ {"id": "nvidia-tensorrt-triton_T5_47_31", "domain": "nvidia-tensorrt-triton", "type": "T5_cross_concept", "query": "How does Genai Perf Analyzer relate to Triton Server?", "ground_truth": ["Genai Perf Analyzer", "Triton Server"], "concept_id_a": 47, "concept_id_b": 31, "hop_depth": 1}
133
+ {"id": "nvidia-tensorrt-triton_T5_38_31", "domain": "nvidia-tensorrt-triton", "type": "T5_cross_concept", "query": "How does Sequence Batching relate to Triton Server?", "ground_truth": ["Sequence Batching", "Triton Server"], "concept_id_a": 38, "concept_id_b": 31, "hop_depth": 1}
134
+ {"id": "nvidia-tensorrt-triton_T5_10_2", "domain": "nvidia-tensorrt-triton", "type": "T5_cross_concept", "query": "How does Global Perf Tuner relate to Builder System?", "ground_truth": ["Global Perf Tuner", "Builder System"], "concept_id_a": 10, "concept_id_b": 2, "hop_depth": 1}
135
+ {"id": "nvidia-tensorrt-triton_T5_35_31", "domain": "nvidia-tensorrt-triton", "type": "T5_cross_concept", "query": "How does Triton C Api relate to Triton Server?", "ground_truth": ["Triton C Api", "Triton Server"], "concept_id_a": 35, "concept_id_b": 31, "hop_depth": 1}
136
+ {"id": "nvidia-tensorrt-triton_T5_15_16", "domain": "nvidia-tensorrt-triton", "type": "T5_cross_concept", "query": "How does Dynamic Shapes relate to Packed Nhd Layout?", "ground_truth": ["Dynamic Shapes", "Packed Nhd Layout"], "concept_id_a": 15, "concept_id_b": 16, "hop_depth": 1}
137
+ {"id": "nvidia-tensorrt-triton_T5_23_9", "domain": "nvidia-tensorrt-triton", "type": "T5_cross_concept", "query": "How does Fused Attention relate to Layer Fusion?", "ground_truth": ["Fused Attention", "Layer Fusion"], "concept_id_a": 23, "concept_id_b": 9, "hop_depth": 1}
138
+ {"id": "nvidia-tensorrt-triton_T5_40_39", "domain": "nvidia-tensorrt-triton", "type": "T5_cross_concept", "query": "How does Bls relate to Ensemble Model?", "ground_truth": ["Bls", "Ensemble Model"], "concept_id_a": 40, "concept_id_b": 39, "hop_depth": 1}
139
+ {"id": "nvidia-tensorrt-triton_T5_21_1", "domain": "nvidia-tensorrt-triton", "type": "T5_cross_concept", "query": "How does Moe Blackwell relate to Tensorrt Sdk?", "ground_truth": ["Moe Blackwell", "Tensorrt Sdk"], "concept_id_a": 21, "concept_id_b": 1, "hop_depth": 1}
140
+ {"id": "nvidia-tensorrt-triton_T5_36_31", "domain": "nvidia-tensorrt-triton", "type": "T5_cross_concept", "query": "How does Backend C Api relate to Triton Server?", "ground_truth": ["Backend C Api", "Triton Server"], "concept_id_a": 36, "concept_id_b": 31, "hop_depth": 1}
141
+ {"id": "nvidia-tensorrt-triton_T5_43_36", "domain": "nvidia-tensorrt-triton", "type": "T5_cross_concept", "query": "How does Onnxrt Backend relate to Backend C Api?", "ground_truth": ["Onnxrt Backend", "Backend C Api"], "concept_id_a": 43, "concept_id_b": 36, "hop_depth": 1}
142
+ {"id": "nvidia-tensorrt-triton_T5_3_2", "domain": "nvidia-tensorrt-triton", "type": "T5_cross_concept", "query": "How does Runtime Engine relate to Builder System?", "ground_truth": ["Runtime Engine", "Builder System"], "concept_id_a": 3, "concept_id_b": 2, "hop_depth": 1}
143
+ {"id": "nvidia-tensorrt-triton_T5_19_6", "domain": "nvidia-tensorrt-triton", "type": "T5_cross_concept", "query": "How does Explicit Quantization relate to Strong Typing Model?", "ground_truth": ["Explicit Quantization", "Strong Typing Model"], "concept_id_a": 19, "concept_id_b": 6, "hop_depth": 1}
144
+ {"id": "nvidia-tensorrt-triton_T5_33_44", "domain": "nvidia-tensorrt-triton", "type": "T5_cross_concept", "query": "How does Http Rest Api relate to Kserve Spec?", "ground_truth": ["Http Rest Api", "Kserve Spec"], "concept_id_a": 33, "concept_id_b": 44, "hop_depth": 1}
145
+ {"id": "nvidia-tensorrt-triton_T5_41_20", "domain": "nvidia-tensorrt-triton", "type": "T5_cross_concept", "query": "How does Trt Llm Backend relate to Kv Cache?", "ground_truth": ["Trt Llm Backend", "Kv Cache"], "concept_id_a": 41, "concept_id_b": 20, "hop_depth": 1}
146
+ {"id": "nvidia-tensorrt-triton_T5_46_31", "domain": "nvidia-tensorrt-triton", "type": "T5_cross_concept", "query": "How does Model Analyzer relate to Triton Server?", "ground_truth": ["Model Analyzer", "Triton Server"], "concept_id_a": 46, "concept_id_b": 31, "hop_depth": 1}
147
+ {"id": "nvidia-tensorrt-triton_T5_43_28", "domain": "nvidia-tensorrt-triton", "type": "T5_cross_concept", "query": "How does Onnxrt Backend relate to Onnx Import?", "ground_truth": ["Onnxrt Backend", "Onnx Import"], "concept_id_a": 43, "concept_id_b": 28, "hop_depth": 1}
148
+ {"id": "nvidia-tensorrt-triton_T5_4_3", "domain": "nvidia-tensorrt-triton", "type": "T5_cross_concept", "query": "How does Trt Cpp Api relate to Runtime Engine?", "ground_truth": ["Trt Cpp Api", "Runtime Engine", "Builder System"], "concept_id_a": 4, "concept_id_b": 3, "hop_depth": 1}
149
+ {"id": "nvidia-tensorrt-triton_T5_2_28", "domain": "nvidia-tensorrt-triton", "type": "T5_cross_concept", "query": "How does Builder System relate to Onnx Import?", "ground_truth": ["Builder System", "Onnx Import"], "concept_id_a": 2, "concept_id_b": 28, "hop_depth": 1}
150
+ {"id": "nvidia-tensorrt-triton_T5_27_1", "domain": "nvidia-tensorrt-triton", "type": "T5_cross_concept", "query": "How does Dla relate to Tensorrt Sdk?", "ground_truth": ["Dla", "Tensorrt Sdk"], "concept_id_a": 27, "concept_id_b": 1, "hop_depth": 1}
151
+ {"id": "nvidia-tensorrt-triton_T5_23_22", "domain": "nvidia-tensorrt-triton", "type": "T5_cross_concept", "query": "How does Fused Attention relate to Rotary Pos Embed?", "ground_truth": ["Fused Attention", "Rotary Pos Embed"], "concept_id_a": 23, "concept_id_b": 22, "hop_depth": 1}
152
+ {"id": "nvidia-tensorrt-triton_T5_1_3", "domain": "nvidia-tensorrt-triton", "type": "T5_cross_concept", "query": "How does Tensorrt Sdk relate to Runtime Engine?", "ground_truth": ["Tensorrt Sdk", "Runtime Engine", "Builder System"], "concept_id_a": 1, "concept_id_b": 3, "hop_depth": 1}
153
+ {"id": "nvidia-tensorrt-triton_T5_4_1", "domain": "nvidia-tensorrt-triton", "type": "T5_cross_concept", "query": "How does Trt Cpp Api relate to Tensorrt Sdk?", "ground_truth": ["Trt Cpp Api", "Tensorrt Sdk", "Builder System", "Runtime Engine"], "concept_id_a": 4, "concept_id_b": 1, "hop_depth": 1}
154
+ {"id": "nvidia-tensorrt-triton_T5_31_34", "domain": "nvidia-tensorrt-triton", "type": "T5_cross_concept", "query": "How does Triton Server relate to Grpc Api?", "ground_truth": ["Triton Server", "Grpc Api"], "concept_id_a": 31, "concept_id_b": 34, "hop_depth": 1}
155
+ {"id": "nvidia-tensorrt-triton_T5_9_2", "domain": "nvidia-tensorrt-triton", "type": "T5_cross_concept", "query": "How does Layer Fusion relate to Builder System?", "ground_truth": ["Layer Fusion", "Builder System"], "concept_id_a": 9, "concept_id_b": 2, "hop_depth": 1}
156
+ {"id": "nvidia-tensorrt-triton_T5_5_3", "domain": "nvidia-tensorrt-triton", "type": "T5_cross_concept", "query": "How does Trt Python Api relate to Runtime Engine?", "ground_truth": ["Trt Python Api", "Runtime Engine", "Builder System"], "concept_id_a": 5, "concept_id_b": 3, "hop_depth": 1}
157
+ {"id": "nvidia-tensorrt-triton_T5_17_8", "domain": "nvidia-tensorrt-triton", "type": "T5_cross_concept", "query": "How does Ptq relate to Precision Modes?", "ground_truth": ["Ptq", "Precision Modes"], "concept_id_a": 17, "concept_id_b": 8, "hop_depth": 1}
158
+ {"id": "nvidia-tensorrt-triton_T5_42_36", "domain": "nvidia-tensorrt-triton", "type": "T5_cross_concept", "query": "How does Vllm Backend relate to Backend C Api?", "ground_truth": ["Vllm Backend", "Backend C Api"], "concept_id_a": 42, "concept_id_b": 36, "hop_depth": 1}
159
+ {"id": "nvidia-tensorrt-triton_T5_20_23", "domain": "nvidia-tensorrt-triton", "type": "T5_cross_concept", "query": "How does Kv Cache relate to Fused Attention?", "ground_truth": ["Kv Cache", "Fused Attention"], "concept_id_a": 20, "concept_id_b": 23, "hop_depth": 1}
160
+ {"id": "nvidia-tensorrt-triton_T5_7_1", "domain": "nvidia-tensorrt-triton", "type": "T5_cross_concept", "query": "How does Ipluginv3 relate to Tensorrt Sdk?", "ground_truth": ["Ipluginv3", "Tensorrt Sdk"], "concept_id_a": 7, "concept_id_b": 1, "hop_depth": 1}
161
+ {"id": "nvidia-tensorrt-triton_T5_17_2", "domain": "nvidia-tensorrt-triton", "type": "T5_cross_concept", "query": "How does Ptq relate to Builder System?", "ground_truth": ["Ptq", "Builder System"], "concept_id_a": 17, "concept_id_b": 2, "hop_depth": 1}
162
+ {"id": "nvidia-tensorrt-triton_T5_34_44", "domain": "nvidia-tensorrt-triton", "type": "T5_cross_concept", "query": "How does Grpc Api relate to Kserve Spec?", "ground_truth": ["Grpc Api", "Kserve Spec"], "concept_id_a": 34, "concept_id_b": 44, "hop_depth": 1}
163
+ {"id": "nvidia-tensorrt-triton_T5_31_32", "domain": "nvidia-tensorrt-triton", "type": "T5_cross_concept", "query": "How does Triton Server relate to Model Repository?", "ground_truth": ["Triton Server", "Model Repository"], "concept_id_a": 31, "concept_id_b": 32, "hop_depth": 1}
164
+ {"id": "nvidia-tensorrt-triton_T5_50_42", "domain": "nvidia-tensorrt-triton", "type": "T5_cross_concept", "query": "How does Speculative Decoding relate to Vllm Backend?", "ground_truth": ["Speculative Decoding", "Vllm Backend"], "concept_id_a": 50, "concept_id_b": 42, "hop_depth": 1}
165
+ {"id": "nvidia-tensorrt-triton_T5_33_31", "domain": "nvidia-tensorrt-triton", "type": "T5_cross_concept", "query": "How does Http Rest Api relate to Triton Server?", "ground_truth": ["Http Rest Api", "Triton Server"], "concept_id_a": 33, "concept_id_b": 31, "hop_depth": 1}
166
+ {"id": "nvidia-tensorrt-triton_T5_24_16", "domain": "nvidia-tensorrt-triton", "type": "T5_cross_concept", "query": "How does Iattention Layer relate to Packed Nhd Layout?", "ground_truth": ["Iattention Layer", "Packed Nhd Layout"], "concept_id_a": 24, "concept_id_b": 16, "hop_depth": 1}
167
+ {"id": "nvidia-tensorrt-triton_T5_10_11", "domain": "nvidia-tensorrt-triton", "type": "T5_cross_concept", "query": "How does Global Perf Tuner relate to Timing Cache?", "ground_truth": ["Global Perf Tuner", "Timing Cache", "Builder System"], "concept_id_a": 10, "concept_id_b": 11, "hop_depth": 1}