Files
sglang/python/sglang/multimodal_gen/test/server/perf_baselines/5090.json
T

129 lines
4.0 KiB
JSON

{
"metadata": {
"model": "Diffusion Server",
"hardware": "CI RTX 5090 pool",
"description": "Reference numbers tightened from RTX 5090 multimodal_gen CI run 31935214180 at source revision 2ee0d38a8553d186faad9c50893480e3c1d79270.",
"last_updated": "2026-08-16"
},
"tolerances": {
"long_term": {
"e2e": 0.15,
"denoise_stage": 0.1,
"non_denoise_stage": 0.5,
"denoise_step": 0.25,
"denoise_agg": 0.15
},
"pr_test": {
"e2e": 0.25,
"denoise_stage": 0.25,
"non_denoise_stage": 0.8,
"denoise_step": 0.3,
"denoise_agg": 0.2
}
},
"improvement_reporting": {
"threshold": 0.2
},
"sampling": {
"step_fractions": [
0.0,
0.2,
0.4,
0.6,
0.8,
1.0
]
},
"scenarios": {
"flux_2_klein_base_image_t2i": {
"stages_ms": {
"DecodingStage": 18.19,
"DenoisingStage": 17612.9,
"ImageVAEEncodingStage": 0.01,
"InputValidationStage": 0.08,
"LatentPreparationStage": 1.67,
"TextEncodingStage": 104.46,
"TimestepPreparationStage": 142.18
},
"denoise_step_ms": {
"0": 95.07,
"10": 344.0,
"20": 345.85,
"29": 346.2,
"39": 346.9,
"49": 347.31
},
"expected_e2e_ms": 18255.39,
"expected_avg_denoise_ms": 338.78,
"expected_median_denoise_ms": 345.95,
"estimated_full_test_time_s": 94.0
},
"wan2_1_t2v_1.3b": {
"stages_ms": {
"DecodingStage": 495.01,
"DenoisingStage": 21452.59,
"InputValidationStage": 0.09,
"LatentPreparationStage": 0.23,
"TextEncodingStage": 1025.11,
"TimestepPreparationStage": 3.91
},
"denoise_step_ms": {
"0": 169.02,
"10": 426.11,
"20": 427.32,
"29": 428.09,
"39": 428.04,
"49": 422.31
},
"expected_e2e_ms": 23418.73,
"expected_avg_denoise_ms": 426.59,
"expected_median_denoise_ms": 427.94,
"estimated_full_test_time_s": 160.9
},
"zimage_image_t2i": {
"stages_ms": {
"DecodingStage": 7.11,
"DenoisingStage": 2178.8,
"InputValidationStage": 0.04,
"LatentPreparationStage": 0.17,
"TextEncodingStage": 252.2,
"TimestepPreparationStage": 36.04
},
"denoise_step_ms": {
"0": 39.34,
"2": 270.3,
"3": 263.6,
"5": 264.28,
"6": 265.1,
"8": 265.56
},
"expected_e2e_ms": 2492.09,
"expected_avg_denoise_ms": 241.53,
"expected_median_denoise_ms": 265.1,
"estimated_full_test_time_s": 329.8
},
"lingbot_video_moe_t2v": {
"stages_ms": {
"InputValidationStage": 0.1,
"LingBotVideoTextEncodingStage": 270.0,
"LatentPreparationStage": 1.2,
"TimestepPreparationStage": 0.2,
"DenoisingStage": 26414.5,
"DecodingStage": 628.3
},
"denoise_step_ms": {
"0": 2200.8,
"2": 2200.8,
"4": 2200.8,
"7": 2200.8,
"9": 2200.8,
"11": 2200.8
},
"expected_e2e_ms": 27740.0,
"expected_avg_denoise_ms": 2200.8,
"expected_median_denoise_ms": 2200.8,
"estimated_full_test_time_s": 90.0
}
}
}