{ "metadata": { "model": "Diffusion Server", "hardware": "CI H100 80GB pool", "description": "Reference numbers captured from the CI diffusion server baseline run" }, "tolerances": { "long_term": { "e2e": 0.1, "denoise_stage": 0.05, "non_denoise_stage": 0.4, "denoise_step": 0.2, "denoise_agg": 0.1 }, "pr_test": { "e2e": 0.15, "denoise_stage": 0.1, "non_denoise_stage": 0.6, "denoise_step": 0.25, "denoise_agg": 0.15 } }, "improvement_reporting": { "threshold": 0.2 }, "sampling": { "step_fractions": [ 0.0, 0.2, 0.4, 0.6, 0.8, 1.0 ] }, "scenarios": { "qwen_image_t2i": { "notes": "Single-image generation using the default prompt", "stages_ms": { "DecodingStage": 51.86, "TextEncodingStage": 611.83, "InputValidationStage": 0.05, "DenoisingStage": 14289.46, "LatentPreparationStage": 0.2, "TimestepPreparationStage": 3.34 }, "denoise_step_ms": { "0": 240.5, "1": 279.1, "2": 283.29, "3": 296.63, "4": 287.72, "5": 283.39, "6": 283.98, "7": 291.82, "8": 283.1, "9": 284.43, "10": 288.95, "11": 285.6, "12": 285.99, "13": 285.47, "14": 289.66, "15": 285.74, "16": 284.15, "17": 290.27, "18": 288.04, "19": 284.57, "20": 286.69, "21": 288.95, "22": 287.09, "23": 285.6, "24": 289.31, "25": 285.48, "26": 285.53, "27": 288.13, "28": 287.65, "29": 285.97, "30": 288.9, "31": 287.97, "32": 286.48, "33": 285.38, "34": 286.62, "35": 288.22, "36": 285.6, "37": 286.61, "38": 287.06, "39": 286.2, "40": 284.6, "41": 285.69, "42": 288.46, "43": 285.53, "44": 285.34, "45": 285.74, "46": 287.25, "47": 285.0, "48": 286.82, "49": 287.19 }, "expected_e2e_ms": 14959.11, "expected_avg_denoise_ms": 285.67, "expected_median_denoise_ms": 286.1 }, "qwen_image_t2i_2_gpus": { "stages_ms": { "InputValidationStage": 0.04, "TextEncodingStage": 693.2, "TimestepPreparationStage": 2.84, "LatentPreparationStage": 9.13, "DenoisingStage": 24529.77, "DecodingStage": 612.79 }, "denoise_step_ms": { "0": 405.94, "1": 420.06, "2": 414.79, "3": 392.4, "4": 408.14, "5": 605.0, "6": 469.39, "7": 574.04, "8": 539.61, "9": 452.93, "10": 279.36, "11": 271.8, "12": 438.26, "13": 552.65, "14": 576.1, "15": 679.84, "16": 543.0, "17": 512.81, "18": 522.27, "19": 545.06, "20": 545.85, "21": 523.83, "22": 519.36, "23": 513.78, "24": 532.54, "25": 524.94, "26": 542.59, "27": 570.91, "28": 568.73, "29": 564.52, "30": 564.57, "31": 544.94, "32": 496.81, "33": 488.98, "34": 457.18, "35": 441.42, "36": 437.44, "37": 477.6, "38": 429.17, "39": 465.55, "40": 448.25, "41": 511.83, "42": 450.6, "43": 375.78, "44": 504.4, "45": 524.44, "46": 535.22, "47": 514.52, "48": 431.58, "49": 410.68 }, "expected_e2e_ms": 25850.45, "expected_avg_denoise_ms": 490.43, "expected_median_denoise_ms": 512.32 }, "flux_image_t2i": { "stages_ms": { "DecodingStage": 32.72, "TextEncodingStage": 51.96, "InputValidationStage": 0.03, "DenoisingStage": 7545.16, "LatentPreparationStage": 0.2, "TimestepPreparationStage": 2.43 }, "denoise_step_ms": { "0": 50.06, "1": 58.88, "2": 151.24, "3": 150.97, "4": 151.23, "5": 151.63, "6": 159.11, "7": 158.31, "8": 153.42, "9": 151.42, "10": 151.91, "11": 151.05, "12": 151.52, "13": 157.2, "14": 152.76, "15": 153.85, "16": 153.02, "17": 151.09, "18": 151.49, "19": 155.13, "20": 155.2, "21": 152.82, "22": 152.2, "23": 150.99, "24": 152.74, "25": 153.45, "26": 153.63, "27": 154.92, "28": 152.72, "29": 151.84, "30": 151.84, "31": 152.44, "32": 153.03, "33": 154.07, "34": 152.36, "35": 153.48, "36": 152.05, "37": 152.45, "38": 152.42, "39": 154.91, "40": 152.68, "41": 153.43, "42": 151.62, "43": 153.52, "44": 153.13, "45": 152.85, "46": 152.33, "47": 151.61, "48": 152.4, "49": 152.33 }, "expected_e2e_ms": 7798.99, "expected_avg_denoise_ms": 150.77, "expected_median_denoise_ms": 152.45 }, "flux_2_image_t2i": { "stages_ms": { "LatentPreparationStage": 0.52, "TimestepPreparationStage": 2.91, "TextEncodingStage": 518.54, "ImageVAEEncodingStage": 0.0, "InputValidationStage": 0.05, "DenoisingStage": 24901.97, "DecodingStage": 8.98 }, "denoise_step_ms": { "0": 69.14, "1": 132.57, "2": 508.67, "3": 493.52, "4": 504.31, "5": 492.99, "6": 501.91, "7": 495.18, "8": 500.87, "9": 497.36, "10": 498.74, "11": 497.46, "12": 499.08, "13": 494.65, "14": 500.35, "15": 496.89, "16": 500.23, "17": 497.01, "18": 501.68, "19": 493.8, "20": 501.1, "21": 494.81, "22": 501.04, "23": 499.27, "24": 500.04, "25": 497.14, "26": 499.05, "27": 494.91, "28": 496.89, "29": 498.53, "30": 497.94, "31": 497.09, "32": 497.7, "33": 497.58, "34": 496.43, "35": 497.7, "36": 497.37, "37": 497.17, "38": 499.27, "39": 495.52, "40": 501.67, "41": 495.11, "42": 500.69, "43": 501.61, "44": 501.91, "45": 495.58, "46": 499.37, "47": 496.8, "48": 497.49, "49": 495.69 }, "expected_e2e_ms": 25832.82, "expected_avg_denoise_ms": 489.43, "expected_median_denoise_ms": 497.53 }, "flux_2_klein_image_t2i": { "stages_ms": { "DecodingStage": 9.27, "TextEncodingStage": 92.17, "InputValidationStage": 0.05, "ImageVAEEncodingStage": 0.0, "DenoisingStage": 252.01, "LatentPreparationStage": 0.42, "TimestepPreparationStage": 1.5 }, "denoise_step_ms": { "0": 19.91, "1": 19.32, "2": 51.99, "3": 61.78 }, "expected_e2e_ms": 430.73, "expected_avg_denoise_ms": 38.25, "expected_median_denoise_ms": 35.95 }, "flux_2_image_t2i_layerwise_offload": { "stages_ms": { "InputValidationStage": 0.06, "TextEncodingStage": 513.58, "ImageVAEEncodingStage": 0.0, "LatentPreparationStage": 0.46, "TimestepPreparationStage": 2.38, "DenoisingStage": 52187.62, "DecodingStage": 190.31 }, "denoise_step_ms": { "0": 1033.45, "1": 137.03, "2": 1046.96, "3": 1039.28, "4": 1039.05, "5": 1043.91, "6": 1041.75, "7": 1037.6, "8": 1043.54, "9": 1048.63, "10": 1039.8, "11": 1042.25, "12": 1041.54, "13": 1045.89, "14": 1038.99, "15": 1041.82, "16": 1038.32, "17": 1045.53, "18": 1046.54, "19": 1041.22, "20": 1044.55, "21": 1041.31, "22": 1051.28, "23": 1043.12, "24": 1044.65, "25": 1042.25, "26": 1046.47, "27": 1052.9, "28": 1039.04, "29": 1042.39, "30": 1045.33, "31": 1038.05, "32": 1037.76, "33": 1037.93, "34": 1052.85, "35": 1045.59, "36": 1054.32, "37": 1044.59, "38": 1043.57, "39": 1041.93, "40": 1043.59, "41": 1046.17, "42": 1046.92, "43": 1047.04, "44": 1046.8, "45": 1041.86, "46": 1041.05, "47": 1044.04, "48": 1039.77, "49": 1047.12 }, "expected_e2e_ms": 53290.15, "expected_avg_denoise_ms": 1025.35, "expected_median_denoise_ms": 1043.33 }, "flux_2_ti2i": { "stages_ms": { "InputValidationStage": 99.82, "TextEncodingStage": 519.88, "ImageVAEEncodingStage": 254.56, "LatentPreparationStage": 12.4, "TimestepPreparationStage": 2.71, "DenoisingStage": 54705.41, "DecodingStage": 469.47 }, "denoise_step_ms": { "0": 1067.03, "1": 271.58, "2": 1073.07, "3": 1071.93, "4": 1100.0, "5": 1102.28, "6": 1088.3, "7": 1089.09, "8": 1086.95, "9": 1089.33, "10": 1089.28, "11": 1096.51, "12": 1098.88, "13": 1080.84, "14": 1098.44, "15": 1100.88, "16": 1086.83, "17": 1090.58, "18": 1096.35, "19": 1086.25, "20": 1082.71, "21": 1097.6, "22": 1098.72, "23": 1100.9, "24": 1099.02, "25": 1101.52, "26": 1098.75, "27": 1101.41, "28": 1091.75, "29": 1087.2, "30": 1101.33, "31": 1098.14, "32": 1100.14, "33": 1098.91, "34": 1100.05, "35": 1099.12, "36": 1100.22, "37": 1103.29, "38": 1092.79, "39": 1086.59, "40": 1094.81, "41": 1105.6, "42": 1100.54, "43": 1099.95, "44": 1096.5, "45": 1086.69, "46": 1095.85, "47": 1092.85, "48": 1086.17, "49": 1099.67 }, "expected_e2e_ms": 56308.23, "expected_avg_denoise_ms": 1077.26, "expected_median_denoise_ms": 1096.5 }, "flux_image_t2i_2_gpus": { "stages_ms": { "InputValidationStage": 0.03, "TextEncodingStage": 74.47, "TimestepPreparationStage": 2.23, "LatentPreparationStage": 6.17, "DenoisingStage": 8400.49, "DecodingStage": 381.56 }, "denoise_step_ms": { "0": 73.27, "1": 166.6, "2": 167.31, "3": 168.7, "4": 168.83, "5": 171.05, "6": 174.64, "7": 170.92, "8": 169.69, "9": 169.21, "10": 167.71, "11": 177.62, "12": 166.44, "13": 174.61, "14": 170.43, "15": 169.47, "16": 167.24, "17": 169.15, "18": 169.51, "19": 172.3, "20": 172.19, "21": 172.36, "22": 168.39, "23": 168.47, "24": 170.55, "25": 170.96, "26": 168.43, "27": 169.01, "28": 169.62, "29": 170.95, "30": 171.83, "31": 171.92, "32": 170.1, "33": 170.46, "34": 169.91, "35": 168.91, "36": 170.27, "37": 170.23, "38": 169.62, "39": 169.66, "40": 169.57, "41": 169.42, "42": 168.59, "43": 171.12, "44": 169.6, "45": 169.93, "46": 171.23, "47": 171.03, "48": 170.14, "49": 169.4 }, "expected_e2e_ms": 9006.3, "expected_avg_denoise_ms": 167.89, "expected_median_denoise_ms": 169.67 }, "zimage_image_t2i": { "stages_ms": { "InputValidationStage": 0.03, "TextEncodingStage": 403.47, "TimestepPreparationStage": 1.41, "LatentPreparationStage": 0.11, "DenoisingStage": 756.21, "DecodingStage": 29.41 }, "denoise_step_ms": { "0": 22.29, "1": 75.04, "2": 93.82, "3": 93.34, "4": 93.38, "5": 93.58, "6": 94.01, "7": 93.97, "8": 94.32 }, "expected_e2e_ms": 1292.92, "expected_avg_denoise_ms": 83.75, "expected_median_denoise_ms": 93.58 }, "zimage_image_t2i_multi_lora": { "stages_ms": { "InputValidationStage": 0.04, "TextEncodingStage": 413.69, "TimestepPreparationStage": 1.3, "LatentPreparationStage": 0.11, "DenoisingStage": 813.7, "DecodingStage": 34.51 }, "denoise_step_ms": { "0": 30.35, "1": 74.53, "2": 99.34, "3": 100.92, "4": 99.46, "5": 100.57, "6": 99.72, "7": 100.86, "8": 103.87 }, "expected_e2e_ms": 1464.31, "expected_avg_denoise_ms": 89.96, "expected_median_denoise_ms": 99.72 }, "zimage_image_t2i_2_gpus": { "stages_ms": { "InputValidationStage": 0.08, "TextEncodingStage": 420.74, "TimestepPreparationStage": 1.5, "LatentPreparationStage": 0.12, "DenoisingStage": 1304.07, "DecodingStage": 37.83 }, "denoise_step_ms": { "0": 49.76, "1": 155.22, "2": 155.98, "3": 156.16, "4": 157.04, "5": 156.54, "6": 156.29, "7": 157.36, "8": 156.05 }, "expected_e2e_ms": 1464.87, "expected_avg_denoise_ms": 144.49, "expected_median_denoise_ms": 156.16 }, "qwen_image_edit_ti2i": { "stages_ms": { "LatentPreparationStage": 0.16, "TimestepPreparationStage": 2.62, "ImageEncodingStage": 1174.26, "ImageVAEEncodingStage": 132.67, "InputValidationStage": 38.1, "DenoisingStage": 38135.64, "DecodingStage": 139.72 }, "denoise_step_ms": { "0": 618.31, "1": 769.07, "2": 766.91, "3": 762.77, "4": 764.26, "5": 765.27, "6": 767.35, "7": 764.18, "8": 766.16, "9": 766.89, "10": 766.1, "11": 764.96, "12": 763.52, "13": 765.22, "14": 765.44, "15": 763.9, "16": 763.19, "17": 764.83, "18": 765.36, "19": 765.19, "20": 765.96, "21": 765.74, "22": 765.87, "23": 764.85, "24": 765.44, "25": 765.95, "26": 766.21, "27": 767.91, "28": 765.45, "29": 764.81, "30": 766.26, "31": 765.37, "32": 766.71, "33": 765.67, "34": 766.64, "35": 765.98, "36": 766.04, "37": 764.19, "38": 765.15, "39": 766.33, "40": 767.68, "41": 765.36, "42": 766.61, "43": 766.06, "44": 765.26, "45": 765.29, "46": 764.64, "47": 766.07, "48": 762.89, "49": 763.01 }, "expected_e2e_ms": 39706.9, "expected_avg_denoise_ms": 762.57, "expected_median_denoise_ms": 765.44 }, "qwen_image_t2i_cache_dit_enabled": { "stages_ms": { "InputValidationStage": 0.05, "TextEncodingStage": 675.95, "TimestepPreparationStage": 3.21, "LatentPreparationStage": 0.2, "DenoisingStage": 5248.83, "DecodingStage": 52.24 }, "denoise_step_ms": { "0": 227.68, "1": 277.41, "2": 276.7, "3": 291.52, "4": 52.8, "5": 6.58, "6": 231.58, "7": 52.55, "8": 7.69, "9": 230.59, "10": 52.58, "11": 7.14, "12": 6.95, "13": 234.71, "14": 53.28, "15": 7.09, "16": 6.63, "17": 233.93, "18": 52.71, "19": 6.64, "20": 6.5, "21": 231.37, "22": 52.28, "23": 6.61, "24": 6.48, "25": 232.86, "26": 54.92, "27": 7.51, "28": 7.19, "29": 233.51, "30": 52.97, "31": 6.72, "32": 7.02, "33": 233.14, "34": 52.47, "35": 6.66, "36": 6.52, "37": 233.84, "38": 51.49, "39": 6.87, "40": 6.74, "41": 233.75, "42": 52.65, "43": 6.62, "44": 6.55, "45": 233.45, "46": 52.33, "47": 6.55, "48": 232.58, "49": 52.84 }, "expected_e2e_ms": 5982.78, "expected_avg_denoise_ms": 104.84, "expected_median_denoise_ms": 102.01 }, "wan2_1_t2v_1.3b_teacache_enabled": { "stages_ms": { "InputValidationStage": 0.07, "TextEncodingStage": 2237.78, "TimestepPreparationStage": 2.1, "LatentPreparationStage": 0.84, "DenoisingStage": 13041.23, "DecodingStage": 1274.63, "per_frame_generation": null }, "denoise_step_ms": { "0": 240.71, "1": 248.13, "2": 246.48, "3": 247.87, "4": 249.38, "5": 246.76, "6": 250.42, "7": 250.81, "8": 250.98, "9": 249.9, "10": 246.72, "11": 249.79, "12": 250.46, "13": 249.19, "14": 247.55, "15": 250.12, "16": 247.57, "17": 247.21, "18": 247.32, "19": 247.42, "20": 248.21, "21": 247.19, "22": 247.72, "23": 247.45, "24": 247.9, "25": 247.87, "26": 247.18, "27": 247.65, "28": 246.91, "29": 248.26, "30": 247.82, "31": 247.73, "32": 247.38, "33": 247.84, "34": 247.46, "35": 247.52, "36": 247.94, "37": 248.76, "38": 248.01, "39": 247.45, "40": 247.84, "41": 248.33, "42": 247.41, "43": 248.16, "44": 248.18, "45": 248.44, "46": 248.65, "47": 247.73, "48": 247.48, "49": 247.54 }, "expected_e2e_ms": 18382.19, "expected_avg_denoise_ms": 260.76, "expected_median_denoise_ms": 247.84 }, "wan2_1_t2v_1.3b": { "stages_ms": { "InputValidationStage": 0.07, "TextEncodingStage": 2237.78, "TimestepPreparationStage": 2.1, "LatentPreparationStage": 0.84, "DenoisingStage": 13041.23, "DecodingStage": 1274.63, "per_frame_generation": null }, "denoise_step_ms": { "0": 224.71, "1": 248.13, "2": 246.48, "3": 247.87, "4": 249.38, "5": 246.76, "6": 250.42, "7": 250.81, "8": 250.98, "9": 249.9, "10": 246.72, "11": 249.79, "12": 250.46, "13": 249.19, "14": 247.55, "15": 250.12, "16": 247.57, "17": 247.21, "18": 247.32, "19": 247.42, "20": 248.21, "21": 247.19, "22": 247.72, "23": 247.45, "24": 247.9, "25": 247.87, "26": 247.18, "27": 247.65, "28": 246.91, "29": 248.26, "30": 247.82, "31": 247.73, "32": 247.38, "33": 247.84, "34": 247.46, "35": 247.52, "36": 247.94, "37": 248.76, "38": 248.01, "39": 247.45, "40": 247.84, "41": 248.33, "42": 247.41, "43": 248.16, "44": 248.18, "45": 248.44, "46": 248.65, "47": 247.73, "48": 247.48, "49": 247.54 }, "expected_e2e_ms": 18382.19, "expected_avg_denoise_ms": 260.76, "expected_median_denoise_ms": 247.84 }, "wan2_1_t2v_1.3b_text_encoder_cpu_offload": { "stages_ms": { "InputValidationStage": 0.09, "TextEncodingStage": 2480.54, "TimestepPreparationStage": 3.73, "LatentPreparationStage": 1.34, "DenoisingStage": 12514.88, "DecodingStage": 1147.6, "per_frame_generation": null }, "denoise_step_ms": { "0": 487.21, "1": 243.47, "2": 244.28, "3": 244.06, "4": 244.77, "5": 245.86, "6": 245.38, "7": 246.74, "8": 246.28, "9": 245.58, "10": 245.6, "11": 245.21, "12": 245.08, "13": 245.03, "14": 245.53, "15": 245.36, "16": 246.17, "17": 245.32, "18": 244.37, "19": 246.83, "20": 245.87, "21": 244.93, "22": 245.11, "23": 245.23, "24": 245.76, "25": 245.44, "26": 246.47, "27": 244.56, "28": 244.76, "29": 244.79, "30": 244.76, "31": 244.8, "32": 245.11, "33": 245.27, "34": 245.37, "35": 245.3, "36": 244.84, "37": 245.26, "38": 245.38, "39": 245.31, "40": 244.7, "41": 245.84, "42": 245.66, "43": 246.68, "44": 245.38, "45": 245.98, "46": 246.02, "47": 245.96, "48": 245.31, "49": 244.99 }, "expected_e2e_ms": 16161.11, "expected_avg_denoise_ms": 250.18, "expected_median_denoise_ms": 245.32 }, "wan2_1_t2v_1.3b_cfg_parallel": { "stages_ms": { "InputValidationStage": 0.08, "TextEncodingStage": 2700.44, "TimestepPreparationStage": 2.82, "LatentPreparationStage": 2.0, "DenoisingStage": 11640.75, "DecodingStage": 890.63, "per_frame_generation": null }, "denoise_step_ms": { "0": 266.91, "1": 211.32, "2": 206.59, "3": 208.12, "4": 210.68, "5": 210.28, "6": 213.92, "7": 211.25, "8": 212.89, "9": 205.35, "10": 205.92, "11": 208.99, "12": 207.1, "13": 208.1, "14": 206.52, "15": 205.5, "16": 205.24, "17": 204.93, "18": 207.05, "19": 203.78, "20": 205.23, "21": 203.87, "22": 204.28, "23": 203.8, "24": 206.02, "25": 207.2, "26": 209.53, "27": 207.46, "28": 206.77, "29": 208.14, "30": 208.05, "31": 208.78, "32": 209.23, "33": 209.72, "34": 208.26, "35": 208.55, "36": 205.24, "37": 204.96, "38": 203.77, "39": 210.2, "40": 202.57, "41": 204.77, "42": 204.96, "43": 203.8, "44": 203.9, "45": 204.49, "46": 207.75, "47": 209.09, "48": 207.51, "49": 207.38 }, "expected_e2e_ms": 15245.6, "expected_avg_denoise_ms": 224.37, "expected_median_denoise_ms": 207.15 }, "turbo_wan2_1_t2v_1.3b": { "stages_ms": { "InputValidationStage": 0.06, "TextEncodingStage": 2508.95, "TimestepPreparationStage": 73.51, "LatentPreparationStage": 1.34, "DmdDenoisingStage": 1285.25, "DecodingStage": 805.04, "per_frame_generation": null }, "denoise_step_ms": { "0": 897.62, "1": 126.04, "2": 126.52, "3": 128.26 }, "expected_e2e_ms": 4686.66, "expected_avg_denoise_ms": 319.61, "expected_median_denoise_ms": 127.39 }, "wan2_2_ti2v_5b": { "stages_ms": { "InputValidationStage": 96.27, "TextEncodingStage": 2238.81, "TimestepPreparationStage": 2.39, "LatentPreparationStage": 27.62, "DenoisingStage": 134069.79, "DecodingStage": 13559.79, "per_frame_generation": null }, "denoise_step_ms": { "0": 3181.0, "1": 2561.67, "2": 2578.49, "3": 2582.1, "4": 2572.24, "5": 2577.72, "6": 2581.35, "7": 2578.79, "8": 2584.98, "9": 2588.49, "10": 2594.37, "11": 2591.19, "12": 2591.32, "13": 2595.35, "14": 2594.35, "15": 2595.62, "16": 2596.35, "17": 2596.11, "18": 2597.24, "19": 2603.13, "20": 2599.9, "21": 2601.48, "22": 2603.58, "23": 2601.13, "24": 2600.47, "25": 2604.13, "26": 2606.04, "27": 2605.3, "28": 2602.02, "29": 2601.83, "30": 2603.57, "31": 2606.63, "32": 2606.1, "33": 2602.24, "34": 2603.29, "35": 2602.34, "36": 2602.16, "37": 2608.14, "38": 2603.48, "39": 2601.7, "40": 2603.96, "41": 2604.58, "42": 2606.67, "43": 2603.52, "44": 2599.88, "45": 2598.66, "46": 2600.74, "47": 2602.31, "48": 2608.4, "49": 2606.02 }, "expected_e2e_ms": 150004.2, "expected_avg_denoise_ms": 2608.84, "expected_median_denoise_ms": 2601.59 }, "qwen_image_edit_2509_ti2i": { "stages_ms": { "InputValidationStage": 213.24, "ImageEncodingStage": 1089.12, "ImageVAEEncodingStage": 304.56, "TimestepPreparationStage": 2.94, "LatentPreparationStage": 0.2, "DenoisingStage": 50724.5, "DecodingStage": 601.02 }, "denoise_step_ms": { "0": 1057.09, "1": 1267.06, "2": 1268.33, "3": 1268.94, "4": 1270.36, "5": 1270.44, "6": 1268.61, "7": 1270.21, "8": 1274.98, "9": 1271.57, "10": 1273.15, "11": 1271.56, "12": 1272.69, "13": 1271.62, "14": 1274.04, "15": 1276.81, "16": 1272.2, "17": 1269.33, "18": 1275.96, "19": 1274.43, "20": 1272.57, "21": 1275.28, "22": 1273.63, "23": 1275.06, "24": 1277.39, "25": 1277.27, "26": 1274.74, "27": 1273.38, "28": 1276.77, "29": 1275.59, "30": 1275.51, "31": 1274.9, "32": 1274.8, "33": 1279.03, "34": 1272.9, "35": 1274.67, "36": 1272.61, "37": 1272.82, "38": 1276.41, "39": 1273.55 }, "expected_e2e_ms": 52938.04, "expected_avg_denoise_ms": 1267.96, "expected_median_denoise_ms": 1273.46 }, "qwen_image_layered_i2i": { "stages_ms": { "QwenImageLayeredBeforeDenoisingStage": 2897.28, "DecodingStage": 312.93, "DenoisingStage": 39417.66, "TimestepPreparationStage": 2.29 }, "denoise_step_ms": { "0": 657.28, "1": 799.2, "2": 790.35, "3": 785.79, "4": 792.9, "5": 795.78, "6": 791.28, "7": 790.87, "8": 786.47, "9": 791.03, "10": 788.77, "11": 790.57, "12": 788.7, "13": 786.01, "14": 791.43, "15": 789.88, "16": 791.18, "17": 792.78, "18": 792.06, "19": 790.47, "20": 792.48, "21": 789.13, "22": 792.12, "23": 789.36, "24": 790.2, "25": 790.87, "26": 792.37, "27": 794.92, "28": 792.9, "29": 791.43, "30": 793.01, "31": 793.71, "32": 794.15, "33": 787.93, "34": 792.12, "35": 794.01, "36": 789.05, "37": 790.51, "38": 793.29, "39": 791.94, "40": 788.94, "41": 788.85, "42": 789.76, "43": 788.89, "44": 791.62, "45": 788.04, "46": 790.03, "47": 786.82, "48": 789.75, "49": 789.0 }, "expected_e2e_ms": 42660.88, "expected_avg_denoise_ms": 788.2, "expected_median_denoise_ms": 790.72 }, "fastwan2_2_ti2v_5b": { "stages_ms": { "InputValidationStage": 300.00, "TextEncodingStage": 843.86, "TimestepPreparationStage": 58.66, "LatentPreparationStage": 28.55, "DmdDenoisingStage": 499.34, "DecodingStage": 1924.01, "per_frame_generation": null }, "denoise_step_ms": { "0": 164.76, "1": 165.6, "2": 165.84 }, "expected_e2e_ms": 7722.91, "expected_avg_denoise_ms": 165.42, "expected_median_denoise_ms": 165.66 }, "fast_hunyuan_video": { "stages_ms": { "InputValidationStage": 0.34, "TextEncodingStage": 550.63, "TimestepPreparationStage": 44.28, "LatentPreparationStage": 0.29, "DenoisingStage": 9154.39, "DecodingStage": 5995.09, "per_frame_generation": null }, "denoise_step_ms": { "0": 485.99, "1": 1399.84, "2": 1399.91, "3": 1397.79, "4": 1400.61, "5": 1402.53 }, "expected_e2e_ms": 16672.15, "expected_avg_denoise_ms": 1608.46, "expected_median_denoise_ms": 1488.48 }, "wan2_2_i2v_a14b_2gpu": { "stages_ms": { "InputValidationStage": 18.45, "TextEncodingStage": 3337.77, "TimestepPreparationStage": 2.9, "LatentPreparationStage": 1.25, "ImageVAEEncodingStage": 1655.89, "DenoisingStage": 106972.82, "DecodingStage": 1355.52, "per_frame_generation": null }, "denoise_step_ms": { "0": 1525.6, "1": 1582.6, "2": 1597.84, "3": 1601.34, "4": 1600.86, "5": 1598.32, "6": 1600.93, "7": 1599.88, "8": 1600.0, "9": 1600.55, "10": 1599.27, "11": 1600.59, "12": 1600.17, "13": 1599.72, "14": 1599.76, "15": 24098.85, "16": 1601.29, "17": 1598.89, "18": 1600.12, "19": 1600.52, "20": 1599.59, "21": 1600.37, "22": 1600.35, "23": 1599.7, "24": 1599.92, "25": 1599.75, "26": 1600.2, "27": 1600.06, "28": 1600.41, "29": 1599.35, "30": 1600.69, "31": 1600.15, "32": 1599.33, "33": 1599.86, "34": 1600.52, "35": 1599.84, "36": 1600.38, "37": 1599.23, "38": 1600.27, "39": 1599.78 }, "expected_e2e_ms": 123182.9887, "expected_avg_denoise_ms": 2831.00, "expected_median_denoise_ms": 1600.09 }, "turbo_wan2_2_i2v_a14b_2gpu": { "stages_ms": { "InputValidationStage": 25.01, "TextEncodingStage": 5198.6, "TimestepPreparationStage": 56.26, "LatentPreparationStage": 1.4, "ImageVAEEncodingStage": 1001.89, "DmdDenoisingStage": 4487.79, "DecodingStage": 821.01, "per_frame_generation": null }, "denoise_step_ms": { "0": 3042.56, "1": 485.88, "2": 721.20, "3": 475.58 }, "expected_e2e_ms": 11605.97, "expected_avg_denoise_ms": 1120.4, "expected_median_denoise_ms": 481.74 }, "wan2_1_i2v_14b_480P_2gpu": { "stages_ms": { "InputValidationStage": 38.23, "TextEncodingStage": 3550.36, "ImageEncodingStage": 3462.55, "TimestepPreparationStage": 2.6, "LatentPreparationStage": 9.73, "ImageVAEEncodingStage": 2290.98, "DenoisingStage": 415021.17, "DecodingStage": 3016.1, "per_frame_generation": null }, "denoise_step_ms": { "0": 10200.25, "1": 8222.39, "2": 8279.38, "3": 8301.48, "4": 8338.87, "5": 8352.39, "6": 8354.64, "7": 8353.64, "8": 8315.58, "9": 8308.48, "10": 8299.65, "11": 8292.7, "12": 8292.73, "13": 8285.21, "14": 8276.06, "15": 8270.41, "16": 8273.04, "17": 8266.04, "18": 8267.7, "19": 8264.06, "20": 8259.32, "21": 8257.26, "22": 8253.02, "23": 8251.77, "24": 8260.97, "25": 8251.39, "26": 8237.43, "27": 8241.33, "28": 8235.96, "29": 8240.6, "30": 8232.48, "31": 8237.85, "32": 8244.3, "33": 8236.79, "34": 8239.83, "35": 8239.89, "36": 8239.12, "37": 8246.74, "38": 8235.67, "39": 8242.77, "40": 8241.17, "41": 8240.24, "42": 8237.01, "43": 8231.26, "44": 8232.85, "45": 8226.56, "46": 8236.98, "47": 8226.73, "48": 8220.49, "49": 8217.04 }, "expected_e2e_ms": 426697.37, "expected_avg_denoise_ms": 8300.19, "expected_median_denoise_ms": 8267.01 }, "wan2_1_i2v_14b_720P_2gpu": { "stages_ms": { "InputValidationStage": 53.67, "TextEncodingStage": 2838, "ImageEncodingStage": 3123.99, "TimestepPreparationStage": 3.39, "LatentPreparationStage": 8.41, "ImageVAEEncodingStage": 2261.05, "DenoisingStage": 417418.12, "DecodingStage": 2968.35 }, "denoise_step_ms": { "0": 11848.08, "1": 8220.3, "2": 8274.3, "3": 8298.9, "4": 8303.34, "5": 8322.44, "6": 8314.37, "7": 8318.54, "8": 8304.94, "9": 8303.04, "10": 8305.22, "11": 8296.22, "12": 8289.2, "13": 8294.19, "14": 8294.87, "15": 8285.96, "16": 8284.98, "17": 8281.61, "18": 8277.35, "19": 8287.46, "20": 8280.3, "21": 8279.18, "22": 8279.37, "23": 8280.16, "24": 8282.67, "25": 8272.14, "26": 8279.37, "27": 8271.66, "28": 8274.6, "29": 8272.88, "30": 8273.76, "31": 8266.17, "32": 8267.77, "33": 8266.88, "34": 8263.14, "35": 8265.97, "36": 8267.76, "37": 8268.03, "38": 8262.24, "39": 8261.4, "40": 8263.65, "41": 8272.46, "42": 8254.9, "43": 8261.03, "44": 8252.92, "45": 8262.49, "46": 8253.67, "47": 8254.92, "48": 8257.08, "49": 8236.56 }, "expected_e2e_ms": 427536.9, "expected_avg_denoise_ms": 8348.21, "expected_median_denoise_ms": 8274.45 }, "wan2_2_t2v_a14b_2gpu": { "stages_ms": { "InputValidationStage": 0.07, "TextEncodingStage": 2575.3, "TimestepPreparationStage": 1.99, "LatentPreparationStage": 1.26, "DenoisingStage": 156678.8406, "DecodingStage": 2702.7, "per_frame_generation": null }, "denoise_step_ms": { "0": 17908.3, "1": 2379.69, "2": 2393.59, "3": 2400.91, "4": 2398.76, "5": 2403.1, "6": 2403.26, "7": 2399.48, "8": 2401.33, "9": 2398.4, "10": 2401.14, "11": 2409.1, "12": 2401.16, "13": 2408.74, "14": 2404.97, "15": 2400.51, "16": 2402.84, "17": 2401.87, "18": 2399.67, "19": 2400.71, "20": 2399.23, "21": 2400.13, "22": 2400.64, "23": 2399.15, "24": 2399.58, "25": 2400.26, "26": 35247.02, "27": 2390.25, "28": 2398.42, "29": 2399.8, "30": 2400.08, "31": 2400.58, "32": 2403.68, "33": 2399.37, "34": 2401.53, "35": 2399.69, "36": 2399.9, "37": 2400.75, "38": 2398.97, "39": 2399.12 }, "expected_e2e_ms": 149864.99, "expected_avg_denoise_ms": 3608.89, "expected_median_denoise_ms": 2400.38 }, "wan2_1_t2v_14b_2gpu": { "stages_ms": { "InputValidationStage": 0.05, "TextEncodingStage": 2310.34, "TimestepPreparationStage": 2.42, "LatentPreparationStage": 27.7, "DenoisingStage": 803631.52, "DecodingStage": 8898.74, "per_frame_generation": null }, "denoise_step_ms": { "0": 17347.88, "1": 15956.93, "2": 16027.54, "3": 16054.15, "4": 16081.46, "5": 16062.7, "6": 16058.56, "7": 16057.58, "8": 16061.04, "9": 16120.97, "10": 16036.84, "11": 16019.6, "12": 16042.29, "13": 16039.87, "14": 16063.0, "15": 16036.16, "16": 16079.82, "17": 16019.7, "18": 16061.5, "19": 16039.95, "20": 16009.42, "21": 16051.01, "22": 16039.31, "23": 16048.22, "24": 16071.41, "25": 16078.75, "26": 16061.78, "27": 16018.39, "28": 16041.44, "29": 16039.64, "30": 16041.89, "31": 16039.6, "32": 16038.97, "33": 15999.48, "34": 16019.93, "35": 16040.27, "36": 16020.3, "37": 16039.38, "38": 15999.4, "39": 16022.15, "40": 16042.32, "41": 16016.62, "42": 15998.92, "43": 16041.48, "44": 15999.63, "45": 16003.21, "46": 15995.91, "47": 16023.52, "48": 16016.64, "49": 16019.6 }, "expected_e2e_ms": 814884.71, "expected_avg_denoise_ms": 16062.92, "expected_median_denoise_ms": 16039.62 }, "wan2_2_t2v_a14b_lora_2gpu": { "stages_ms": { "InputValidationStage": 0.09, "TextEncodingStage": 2552.97, "TimestepPreparationStage": 1.99, "LatentPreparationStage": 1.29, "DenoisingStage": 154340.69, "DecodingStage": 2730.86, "per_frame_generation": null }, "denoise_step_ms": { "0": 26510.7, "1": 2381.25, "2": 2396.9, "3": 2400.96, "4": 2402.47, "5": 2399.6, "6": 2400.5, "7": 2401.13, "8": 2399.32, "9": 2400.0, "10": 2401.35, "11": 2400.04, "12": 2408.27, "13": 2407.08, "14": 2405.92, "15": 2403.99, "16": 2402.12, "17": 2402.52, "18": 2398.08, "19": 2399.9, "20": 2400.14, "21": 2398.64, "22": 2401.32, "23": 2400.75, "24": 2399.27, "25": 2400.21, "26": 36387.55, "27": 2399.77, "28": 2398.09, "29": 2404.64, "30": 2400.68, "31": 2404.3, "32": 2392.44, "33": 2390.56, "34": 2396.05, "35": 2394.86, "36": 2396.07, "37": 2398.49, "38": 2394.77, "39": 2394.19 }, "expected_e2e_ms": 159643.06, "expected_avg_denoise_ms": 3851.87, "expected_median_denoise_ms": 2400.09 }, "wan2_1_t2v_1_3b_lora_1gpu": { "stages_ms": { "InputValidationStage": 0.06, "TextEncodingStage": 2467.44, "TimestepPreparationStage": 2.96, "LatentPreparationStage": 1.87, "DenoisingStage": 14859.47, "DecodingStage": 1199.31, "per_frame_generation": null }, "denoise_step_ms": { "0": 233.29, "1": 265.02, "2": 257.83, "3": 260.27, "4": 261.43, "5": 258.58, "6": 256.64, "7": 256.91, "8": 258.41, "9": 257.84, "10": 257.08, "11": 257.0, "12": 258.44, "13": 257.1, "14": 256.95, "15": 257.2, "16": 256.84, "17": 257.64, "18": 257.22, "19": 257.42, "20": 256.91, "21": 256.99, "22": 257.17, "23": 257.63, "24": 258.89, "25": 257.46, "26": 257.3, "27": 257.42, "28": 257.19, "29": 257.65, "30": 257.39, "31": 256.93, "32": 258.23, "33": 257.62, "34": 281.86, "35": 295.86, "36": 296.73, "37": 287.21, "38": 300.87, "39": 303.47, "40": 294.09, "41": 270.52, "42": 256.53, "43": 256.58, "44": 256.29, "45": 255.81, "46": 256.34, "47": 256.08, "48": 255.92, "49": 255.87 }, "expected_e2e_ms": 18547.46, "expected_avg_denoise_ms": 297.09, "expected_median_denoise_ms": 257.42 }, "wan2_1_i2v_14b_lora_2gpu": { "stages_ms": { "InputValidationStage": 23.97, "TextEncodingStage": 2485.39, "ImageEncodingStage": 2372.07, "TimestepPreparationStage": 2.6, "LatentPreparationStage": 0.18, "ImageVAEEncodingStage": 2500.13, "DenoisingStage": 193514.04, "DecodingStage": 3341.78, "per_frame_generation": null }, "denoise_step_ms": { "0": 6680.62, "1": 3765.8, "2": 3774.63, "3": 3772.93, "4": 3781.13, "5": 3778.22, "6": 3776.41, "7": 3772.02, "8": 3776.15, "9": 3768.82, "10": 3775.31, "11": 3771.32, "12": 3774.33, "13": 3772.5, "14": 3778.41, "15": 3775.31, "16": 3771.38, "17": 3774.87, "18": 3780.01, "19": 3772.85, "20": 3773.65, "21": 3774.47, "22": 3774.39, "23": 3773.08, "24": 3776.71, "25": 3780.01, "26": 3774.83, "27": 3773.27, "28": 3773.76, "29": 3772.75, "30": 3773.01, "31": 3773.34, "32": 3773.13, "33": 3774.12, "34": 3772.19, "35": 3774.7, "36": 3773.98, "37": 3772.47, "38": 3771.72, "39": 3774.07, "40": 3773.71, "41": 3773.6, "42": 3772.12, "43": 3773.75, "44": 3782.43, "45": 3779.66, "46": 3779.86, "47": 3774.58, "48": 3770.54, "49": 3776.76 }, "expected_e2e_ms": 204257.12, "expected_avg_denoise_ms": 3855.55, "expected_median_denoise_ms": 3774.03 }, "flux_2_image_t2i_2_gpus": { "stages_ms": { "InputValidationStage": 0.05, "TextEncodingStage": 518.88, "ImageVAEEncodingStage": 0.0, "LatentPreparationStage": 0.45, "TimestepPreparationStage": 3.41, "DenoisingStage": 26377.63, "DecodingStage": 321.94 }, "denoise_step_ms": { "0": 129.07, "1": 437.16, "2": 437.7, "3": 437.67, "4": 437.84, "5": 438.03, "6": 438.09, "7": 437.65, "8": 437.95, "9": 438.31, "10": 437.99, "11": 438.54, "12": 438.47, "13": 438.2, "14": 438.56, "15": 438.69, "16": 438.69, "17": 438.98, "18": 437.96, "19": 438.9, "20": 438.87, "21": 438.04, "22": 437.88, "23": 439.09, "24": 438.61, "25": 437.68, "26": 439.2, "27": 439.63, "28": 438.65, "29": 439.32, "30": 439.01, "31": 438.84, "32": 438.72, "33": 439.09, "34": 438.3, "35": 439.48, "36": 438.2, "37": 439.67, "38": 440.65, "39": 439.96, "40": 439.0, "41": 439.2, "42": 439.37, "43": 439.98, "44": 438.6, "45": 439.58, "46": 440.23, "47": 440.1, "48": 440.21, "49": 439.22 }, "expected_e2e_ms": 27624.8, "expected_avg_denoise_ms": 518.23, "expected_median_denoise_ms": 528.06 }, "qwen_image_edit_2511_ti2i": { "stages_ms": { "InputValidationStage": 55.15, "ImageEncodingStage": 770.33, "ImageVAEEncodingStage": 88.06, "TimestepPreparationStage": 2.12, "LatentPreparationStage": 0.14, "DenoisingStage": 23869.32, "DecodingStage": 108.23 }, "denoise_step_ms": { "0": 478.35, "1": 608.56, "2": 588.51, "3": 607.26, "4": 599.37, "5": 595.19, "6": 603.22, "7": 594.48, "8": 605.06, "9": 597.63, "10": 601.03, "11": 597.18, "12": 598.82, "13": 600.05, "14": 598.57, "15": 601.4, "16": 595.17, "17": 599.21, "18": 600.86, "19": 600.93, "20": 600.35, "21": 600.63, "22": 597.58, "23": 600.73, "24": 599.36, "25": 600.48, "26": 600.33, "27": 599.34, "28": 599.61, "29": 599.71, "30": 596.03, "31": 599.85, "32": 599.36, "33": 601.58, "34": 597.91, "35": 600.79, "36": 599.29, "37": 601.64, "38": 598.24, "39": 599.87 }, "expected_e2e_ms": 24895.28, "expected_avg_denoise_ms": 596.59, "expected_median_denoise_ms": 599.66 }, "fsdp-inference": { "stages_ms": { "InputValidationStage": 0.04, "TextEncodingStage": 411.12, "TimestepPreparationStage": 1.44, "LatentPreparationStage": 0.1, "DenoisingStage": 1569.61, "DecodingStage": 41.43 }, "denoise_step_ms": { "0": 165.33, "1": 158.34, "2": 167.65, "3": 179.11, "4": 183.98, "5": 175.08, "6": 178.34, "7": 178.53, "8": 178.08 }, "expected_e2e_ms": 2103.05, "expected_avg_denoise_ms": 173.83, "expected_median_denoise_ms": 178.08 } } }