From 1456eb612d395206bfe4798498e6a60d8173e411 Mon Sep 17 00:00:00 2001 From: Mick Date: Sun, 14 Jun 2026 15:50:21 +0800 Subject: [PATCH] [diffusion] CI: tighten perf baselines (#28123) --- .../test/server/perf_baselines.json | 2726 ++++++++--------- 1 file changed, 1363 insertions(+), 1363 deletions(-) diff --git a/python/sglang/multimodal_gen/test/server/perf_baselines.json b/python/sglang/multimodal_gen/test/server/perf_baselines.json index 99618d9f6..677c90dda 100644 --- a/python/sglang/multimodal_gen/test/server/perf_baselines.json +++ b/python/sglang/multimodal_gen/test/server/perf_baselines.json @@ -2,8 +2,8 @@ "metadata": { "model": "Diffusion Server", "hardware": "CI H100 80GB pool", - "description": "Reference numbers captured from CI H100 runs (2026-05-17).", - "last_updated": "2026-05-17" + "description": "Reference numbers captured from CI diffusion runs (latest source: PR 28127 run 27487958425, tightened only from prior baselines).", + "last_updated": "2026-06-14" }, "tolerances": { "long_term": { @@ -37,36 +37,36 @@ "scenarios": { "qwen_image_t2i": { "stages_ms": { - "TextEncodingStage": 520.0, + "TextEncodingStage": 232.03, "DenoisingStage": 12404.23, "InputValidationStage": 0.05, - "LatentPreparationStage": 0.23, + "LatentPreparationStage": 0.22, "TimestepPreparationStage": 14.99, "DecodingStage": 17.7 }, "denoise_step_ms": { "0": 196.21, "1": 242.25, - "2": 245.88, - "3": 257.46, + "2": 243.58, + "3": 256.87, "4": 249.73, - "5": 245.97, - "6": 246.48, + "5": 244.48, + "6": 244.27, "7": 253.29, "8": 245.72, - "9": 246.87, - "10": 250.41, + "9": 246.67, + "10": 245.49, "11": 247.89, "12": 248.23, - "13": 247.78, - "14": 251.23, + "13": 247.24, + "14": 247.57, "15": 248.01, "16": 246.63, - "17": 251.28, - "18": 250.01, + "17": 248.05, + "18": 248.94, "19": 247.0, "20": 248.84, - "21": 250.8, + "21": 248.0, "22": 249.18, "23": 247.89, "24": 251.11, @@ -94,33 +94,33 @@ "46": 249.32, "47": 247.37, "48": 248.95, - "49": 249.27 + "49": 248.9 }, - "expected_e2e_ms": 13099.2, + "expected_e2e_ms": 12799.17, "expected_avg_denoise_ms": 247.95, "expected_median_denoise_ms": 249.01, "estimated_full_test_time_s": 133.1 }, "qwen_image_t2i_2_gpus": { "stages_ms": { - "InputValidationStage": 0.13, - "TextEncodingStage": 1050.0, - "LatentPreparationStage": 0.26, + "InputValidationStage": 0.04, + "TextEncodingStage": 457.66, + "LatentPreparationStage": 0.21, "TimestepPreparationStage": 20.44, - "DenoisingStage": 10159.11, + "DenoisingStage": 10125.16, "DecodingStage": 13.71 }, "denoise_step_ms": { - "0": 230.15, - "1": 241.11, - "2": 267.78, - "3": 276.94, - "4": 273.46, - "5": 263.83, - "6": 262.48, - "7": 252.42, - "8": 258.84, - "9": 238.48, + "0": 208.21, + "1": 231.64, + "2": 232.46, + "3": 228.42, + "4": 227.95, + "5": 229.55, + "6": 235.94, + "7": 222.3, + "8": 224.02, + "9": 228.53, "10": 197.15, "11": 183.96, "12": 184.16, @@ -131,7 +131,7 @@ "17": 189.98, "18": 190.28, "19": 189.91, - "20": 221.0, + "20": 199.31, "21": 190.14, "22": 190.16, "23": 189.96, @@ -148,22 +148,22 @@ "34": 189.1, "35": 190.04, "36": 189.37, - "37": 192.09, - "38": 189.15, - "39": 307.1, + "37": 187.14, + "38": 188.47, + "39": 188.48, "40": 188.16, - "41": 191.37, + "41": 187.64, "42": 188.23, - "43": 189.88, - "44": 189.97, - "45": 190.1, + "43": 188.18, + "44": 189.71, + "45": 188.61, "46": 189.33, "47": 190.56, "48": 190.14, "49": 189.98 }, - "expected_e2e_ms": 11362.7, - "expected_avg_denoise_ms": 203.06, + "expected_e2e_ms": 10681.51, + "expected_avg_denoise_ms": 202.35, "expected_median_denoise_ms": 190.02, "estimated_full_test_time_s": 133.2 }, @@ -233,231 +233,231 @@ "flux_image_t2i": { "stages_ms": { "TimestepPreparationStage": 32.58, - "DenoisingStage": 6915.27, + "DenoisingStage": 6876.88, "DecodingStage": 10.31, "LatentPreparationStage": 0.19, - "TextEncodingStage": 52.38, - "InputValidationStage": 0.07 + "TextEncodingStage": 22.38, + "InputValidationStage": 0.05 }, "denoise_step_ms": { - "0": 45.85, + "0": 45.19, "1": 53.93, - "2": 138.52, - "3": 137.94, - "4": 138.51, - "5": 137.46, - "6": 139.84, + "2": 135.69, + "3": 136.11, + "4": 135.51, + "5": 136.88, + "6": 138.7, "7": 145.0, - "8": 139.97, - "9": 138.69, - "10": 139.13, - "11": 138.35, - "12": 138.78, - "13": 142.34, + "8": 139.34, + "9": 135.89, + "10": 137.33, + "11": 136.49, + "12": 136.15, + "13": 139.22, "14": 139.91, - "15": 140.68, - "16": 140.15, - "17": 138.38, - "18": 138.75, - "19": 138.28, + "15": 139.9, + "16": 138.58, + "17": 136.39, + "18": 137.38, + "19": 137.25, "20": 139.79, "21": 139.97, "22": 139.4, "23": 138.29, - "24": 139.89, - "25": 140.54, - "26": 140.71, - "27": 141.89, + "24": 138.59, + "25": 137.13, + "26": 136.88, + "27": 140.51, "28": 139.88, "29": 139.07, "30": 139.07, - "31": 139.62, - "32": 140.03, - "33": 141.11, + "31": 138.89, + "32": 138.84, + "33": 137.37, "34": 139.55, "35": 140.57, "36": 139.26, - "37": 139.63, - "38": 139.6, - "39": 140.13, - "40": 139.84, - "41": 140.53, + "37": 139.38, + "38": 138.18, + "39": 137.97, + "40": 137.53, + "41": 138.01, "42": 138.87, "43": 140.05, - "44": 140.25, - "45": 140.0, - "46": 139.52, - "47": 138.86, - "48": 139.58, - "49": 139.52 + "44": 139.59, + "45": 139.68, + "46": 139.42, + "47": 137.81, + "48": 138.16, + "49": 139.34 }, - "expected_e2e_ms": 7351.97, - "expected_avg_denoise_ms": 138.09, - "expected_median_denoise_ms": 138.97, + "expected_e2e_ms": 7152.96, + "expected_avg_denoise_ms": 137.39, + "expected_median_denoise_ms": 138.77, "estimated_full_test_time_s": 127.4 }, "flux_2_image_t2i": { "stages_ms": { "TimestepPreparationStage": 20.47, - "DenoisingStage": 24036.73, - "DecodingStage": 11.76, + "DenoisingStage": 23650.96, + "DecodingStage": 7.1, "LatentPreparationStage": 1.17, - "TextEncodingStage": 430.0, - "InputValidationStage": 0.05, + "TextEncodingStage": 375.31, + "InputValidationStage": 0.04, "ImageVAEEncodingStage": 0.01 }, "denoise_step_ms": { "0": 65.28, "1": 129.11, - "2": 494.48, - "3": 472.21, - "4": 485.49, - "5": 477.0, - "6": 484.63, - "7": 481.51, - "8": 476.77, - "9": 484.36, - "10": 474.3, - "11": 484.46, - "12": 472.48, + "2": 488.44, + "3": 465.17, + "4": 478.82, + "5": 472.84, + "6": 475.45, + "7": 477.05, + "8": 469.78, + "9": 480.78, + "10": 466.21, + "11": 482.63, + "12": 468.39, "13": 481.72, - "14": 476.01, - "15": 483.91, - "16": 474.13, - "17": 484.02, - "18": 477.31, - "19": 480.9, - "20": 476.66, - "21": 481.88, - "22": 478.32, - "23": 479.76, - "24": 479.35, - "25": 480.63, - "26": 481.81, - "27": 481.98, - "28": 479.82, - "29": 482.68, - "30": 483.31, - "31": 479.62, - "32": 483.89, - "33": 476.98, - "34": 483.46, - "35": 481.28, - "36": 482.2, - "37": 480.9, - "38": 484.39, - "39": 478.42, - "40": 484.07, - "41": 478.36, - "42": 482.26, - "43": 483.16, - "44": 484.7, - "45": 482.63, - "46": 482.24, - "47": 481.59, - "48": 479.8, - "49": 480.94 + "14": 468.31, + "15": 480.25, + "16": 468.87, + "17": 480.89, + "18": 471.87, + "19": 477.75, + "20": 470.42, + "21": 476.72, + "22": 475.24, + "23": 475.18, + "24": 477.48, + "25": 474.97, + "26": 477.79, + "27": 474.04, + "28": 479.14, + "29": 473.3, + "30": 478.18, + "31": 473.57, + "32": 479.55, + "33": 474.66, + "34": 476.27, + "35": 474.9, + "36": 476.92, + "37": 475.12, + "38": 477.88, + "39": 474.87, + "40": 476.03, + "41": 477.65, + "42": 475.38, + "43": 477.2, + "44": 473.63, + "45": 477.37, + "46": 472.12, + "47": 477.77, + "48": 476.58, + "49": 476.52 }, - "expected_e2e_ms": 24999.55, - "expected_avg_denoise_ms": 472.44, - "expected_median_denoise_ms": 481.55, + "expected_e2e_ms": 24561.42, + "expected_avg_denoise_ms": 464.89, + "expected_median_denoise_ms": 472.74, "estimated_full_test_time_s": 145.2 }, "flux_2_klein_image_t2i": { "stages_ms": { "DecodingStage": 6.41, - "InputValidationStage": 0.05, + "InputValidationStage": 0.03, "DenoisingStage": 266.12, - "TextEncodingStage": 90.44, + "TextEncodingStage": 53.59, "LatentPreparationStage": 0.5, - "TimestepPreparationStage": 20.45, + "TimestepPreparationStage": 19.97, "ImageVAEEncodingStage": 0.01 }, "denoise_step_ms": { "0": 16.21, - "1": 15.23, - "2": 54.97, - "3": 65.32 + "1": 13.42, + "2": 17.64, + "3": 63.6 }, - "expected_e2e_ms": 476.48, + "expected_e2e_ms": 434.57, "expected_avg_denoise_ms": 39.98, "expected_median_denoise_ms": 39.47, "estimated_full_test_time_s": 120.5 }, "flux_2_klein_base_image_t2i": { "stages_ms": { - "InputValidationStage": 0.06, - "TextEncodingStage": 505.51, + "InputValidationStage": 0.03, + "TextEncodingStage": 53.32, "ImageVAEEncodingStage": 0.01, "LatentPreparationStage": 1.14, - "TimestepPreparationStage": 53.76, - "DenoisingStage": 12365.93, - "DecodingStage": 11.74 + "TimestepPreparationStage": 20.37, + "DenoisingStage": 6704.85, + "DecodingStage": 7.23 }, "denoise_step_ms": { "0": 72.0, - "1": 192.12, - "2": 217.22, - "3": 217.5, - "4": 257.12, - "5": 262.19, - "6": 246.18, - "7": 224.94, - "8": 252.71, - "9": 262.82, - "10": 246.37, - "11": 231.87, - "12": 253.34, - "13": 260.94, - "14": 242.16, - "15": 235.92, - "16": 254.31, - "17": 262.02, - "18": 243.92, - "19": 239.61, - "20": 253.47, - "21": 259.75, - "22": 245.48, - "23": 240.91, - "24": 253.84, - "25": 255.36, - "26": 248.31, - "27": 243.75, - "28": 250.5, - "29": 251.34, - "30": 246.87, - "31": 243.8, - "32": 249.8, - "33": 254.02, - "34": 247.4, - "35": 244.77, - "36": 252.03, - "37": 248.85, - "38": 249.74, - "39": 248.6, - "40": 252.29, - "41": 249.25, - "42": 249.85, - "43": 249.68, - "44": 252.65, - "45": 249.2, - "46": 249.32, - "47": 248.62, - "48": 252.87, - "49": 249.19 + "1": 86.92, + "2": 130.18, + "3": 130.1, + "4": 131.2, + "5": 130.92, + "6": 130.31, + "7": 144.42, + "8": 136.01, + "9": 132.71, + "10": 130.96, + "11": 130.37, + "12": 130.77, + "13": 132.92, + "14": 138.92, + "15": 138.26, + "16": 132.61, + "17": 133.87, + "18": 133.08, + "19": 131.36, + "20": 132.41, + "21": 136.52, + "22": 136.33, + "23": 135.24, + "24": 134.33, + "25": 132.95, + "26": 132.93, + "27": 130.62, + "28": 133.91, + "29": 135.79, + "30": 135.71, + "31": 134.15, + "32": 133.72, + "33": 133.32, + "34": 132.37, + "35": 135.04, + "36": 135.56, + "37": 136.58, + "38": 135.37, + "39": 134.37, + "40": 135.29, + "41": 132.97, + "42": 132.97, + "43": 134.56, + "44": 136.54, + "45": 135.2, + "46": 135.27, + "47": 133.13, + "48": 131.98, + "49": 133.24 }, - "expected_e2e_ms": 13075.51, - "expected_avg_denoise_ms": 243.34, - "expected_median_denoise_ms": 249.22, + "expected_e2e_ms": 6870.9, + "expected_avg_denoise_ms": 131.92, + "expected_median_denoise_ms": 133.4, "estimated_full_test_time_s": 124.4 }, "layerwise_offload": { "stages_ms": { - "TextEncodingStage": 176.59, + "TextEncodingStage": 170.96, "DenoisingStage": 1478.72, "InputValidationStage": 0.05, "LatentPreparationStage": 0.12, - "TimestepPreparationStage": 41.81, + "TimestepPreparationStage": 39.69, "DecodingStage": 10.36 }, "denoise_step_ms": { @@ -519,69 +519,69 @@ }, "flux_2_ti2i": { "stages_ms": { - "TextEncodingStage": 430.0, - "DenoisingStage": 47133.15, + "TextEncodingStage": 367.56, + "DenoisingStage": 46838.26, "InputValidationStage": 43.89, - "LatentPreparationStage": 1.19, + "LatentPreparationStage": 0.27, "TimestepPreparationStage": 10.54, "DecodingStage": 16.32, "ImageVAEEncodingStage": 69.59 }, "denoise_step_ms": { "0": 910.0, - "1": 236.54, - "2": 934.61, - "3": 933.62, - "4": 932.9, - "5": 936.67, - "6": 938.92, - "7": 940.38, - "8": 943.45, - "9": 941.77, - "10": 943.22, - "11": 944.59, - "12": 947.95, - "13": 941.38, - "14": 942.94, - "15": 945.5, - "16": 946.09, - "17": 945.03, - "18": 946.27, - "19": 946.09, - "20": 943.01, - "21": 945.24, - "22": 946.88, - "23": 945.43, - "24": 945.39, - "25": 945.36, - "26": 945.47, - "27": 944.91, - "28": 944.35, - "29": 944.95, - "30": 945.26, - "31": 945.0, - "32": 941.29, - "33": 942.61, - "34": 942.31, - "35": 942.54, - "36": 941.52, - "37": 943.8, - "38": 940.42, - "39": 943.78, - "40": 943.87, - "41": 941.57, - "42": 940.55, - "43": 941.47, - "44": 941.05, - "45": 941.33, - "46": 941.46, - "47": 941.03, + "1": 148.53, + "2": 932.49, + "3": 930.25, + "4": 928.98, + "5": 928.58, + "6": 933.5, + "7": 934.46, + "8": 934.42, + "9": 935.13, + "10": 934.08, + "11": 934.95, + "12": 934.47, + "13": 935.96, + "14": 935.65, + "15": 937.07, + "16": 936.13, + "17": 932.69, + "18": 938.8, + "19": 937.48, + "20": 939.61, + "21": 938.2, + "22": 936.83, + "23": 939.99, + "24": 935.59, + "25": 934.87, + "26": 937.1, + "27": 939.72, + "28": 938.59, + "29": 938.43, + "30": 939.54, + "31": 938.4, + "32": 936.73, + "33": 936.87, + "34": 936.71, + "35": 937.35, + "36": 937.94, + "37": 936.72, + "38": 936.04, + "39": 937.37, + "40": 937.72, + "41": 939.13, + "42": 938.47, + "43": 939.74, + "44": 939.68, + "45": 939.95, + "46": 940.96, + "47": 939.01, "48": 940.73, - "49": 940.73 + "49": 940.47 }, - "expected_e2e_ms": 48425.56, - "expected_avg_denoise_ms": 926.43, - "expected_median_denoise_ms": 942.95, + "expected_e2e_ms": 47799.39, + "expected_avg_denoise_ms": 920.57, + "expected_median_denoise_ms": 936.97, "estimated_full_test_time_s": 168.9 }, "flux_2_ti2i_multi_image_cache_dit": { @@ -653,82 +653,82 @@ }, "flux_image_t2i_2_gpus": { "stages_ms": { - "InputValidationStage": 0.08, - "TextEncodingStage": 38.63, - "TimestepPreparationStage": 38.66, - "LatentPreparationStage": 0.38, - "DenoisingStage": 5000.66, + "InputValidationStage": 0.04, + "TextEncodingStage": 28.24, + "TimestepPreparationStage": 25.74, + "LatentPreparationStage": 0.25, + "DenoisingStage": 4994.21, "DecodingStage": 11.98 }, "denoise_step_ms": { "0": 73.63, - "1": 88.19, - "2": 100.47, + "1": 81.34, + "2": 99.66, "3": 99.12, - "4": 99.99, - "5": 100.15, + "4": 99.9, + "5": 100.06, "6": 100.02, - "7": 100.4, - "8": 100.26, + "7": 99.58, + "8": 99.56, "9": 100.84, - "10": 100.96, + "10": 100.78, "11": 100.72, - "12": 101.28, - "13": 100.61, - "14": 100.61, - "15": 100.87, - "16": 100.38, + "12": 100.77, + "13": 100.35, + "14": 100.09, + "15": 100.51, + "16": 99.86, "17": 100.29, - "18": 101.28, + "18": 100.72, "19": 99.24, - "20": 100.75, + "20": 100.42, "21": 100.98, "22": 100.47, - "23": 101.18, + "23": 101.11, "24": 100.19, "25": 101.35, - "26": 101.01, - "27": 101.26, - "28": 100.98, + "26": 100.53, + "27": 100.2, + "28": 99.65, "29": 98.85, - "30": 100.95, - "31": 100.97, + "30": 99.6, + "31": 100.72, "32": 100.29, - "33": 100.59, - "34": 101.09, - "35": 102.19, + "33": 100.43, + "34": 101.03, + "35": 100.54, "36": 99.26, - "37": 100.97, + "37": 100.48, "38": 100.88, - "39": 100.53, - "40": 100.84, + "39": 100.52, + "40": 100.48, "41": 100.76, "42": 100.87, - "43": 100.82, - "44": 100.79, + "43": 100.34, + "44": 100.68, "45": 100.73, - "46": 101.13, - "47": 100.73, - "48": 101.11, + "46": 99.13, + "47": 100.54, + "48": 100.57, "49": 100.09 }, - "expected_e2e_ms": 5242.81, - "expected_avg_denoise_ms": 99.84, - "expected_median_denoise_ms": 100.74, + "expected_e2e_ms": 5215.39, + "expected_avg_denoise_ms": 99.72, + "expected_median_denoise_ms": 100.52, "estimated_full_test_time_s": 125.5 }, "zimage_image_t2i": { "stages_ms": { "DecodingStage": 8.86, - "InputValidationStage": 0.05, + "InputValidationStage": 0.04, "DenoisingStage": 675.8, - "TextEncodingStage": 155.0, + "TextEncodingStage": 130.38, "LatentPreparationStage": 0.14, - "TimestepPreparationStage": 36.26 + "TimestepPreparationStage": 31.38 }, "denoise_step_ms": { "0": 19.93, - "1": 37.9, + "1": 28.65, "2": 83.91, "3": 83.48, "4": 83.51, @@ -744,23 +744,23 @@ }, "zimage_image_t2i_fp8": { "stages_ms": { - "TextEncodingStage": 155.0, + "TextEncodingStage": 129.84, "DenoisingStage": 634.42, - "InputValidationStage": 0.04, + "InputValidationStage": 0.03, "LatentPreparationStage": 0.11, "TimestepPreparationStage": 17.42, "DecodingStage": 9.46 }, "denoise_step_ms": { "0": 33.35, - "1": 36.91, + "1": 36.43, "2": 66.88, "3": 78.15, "4": 78.0, "5": 78.36, "6": 78.48, "7": 78.32, - "8": 78.53 + "8": 73.36 }, "expected_e2e_ms": 958.32, "expected_avg_denoise_ms": 70.04, @@ -769,17 +769,17 @@ }, "zimage_image_t2i_multi_lora": { "stages_ms": { - "TimestepPreparationStage": 40.51, + "TimestepPreparationStage": 24.93, "DenoisingStage": 673.95, "DecodingStage": 8.43, - "LatentPreparationStage": 0.11, - "TextEncodingStage": 155.0, - "InputValidationStage": 0.04 + "LatentPreparationStage": 0.1, + "TextEncodingStage": 129.4, + "InputValidationStage": 0.03 }, "denoise_step_ms": { "0": 25.08, "1": 35.42, - "2": 82.08, + "2": 77.56, "3": 83.39, "4": 82.18, "5": 83.1, @@ -794,10 +794,10 @@ }, "zimage_image_t2i_2_gpus": { "stages_ms": { - "InputValidationStage": 0.05, - "TextEncodingStage": 309.68, + "InputValidationStage": 0.04, + "TextEncodingStage": 256.67, "LatentPreparationStage": 0.14, - "TimestepPreparationStage": 37.19, + "TimestepPreparationStage": 36.44, "DenoisingStage": 525.39, "DecodingStage": 10.18 }, @@ -807,9 +807,9 @@ "2": 54.71, "3": 64.29, "4": 65.0, - "5": 65.56, + "5": 63.7, "6": 64.61, - "7": 64.76, + "7": 64.42, "8": 64.35 }, "expected_e2e_ms": 961.86, @@ -819,272 +819,272 @@ }, "qwen_image_edit_ti2i": { "stages_ms": { - "InputValidationStage": 31.6, - "ImageEncodingStage": 1049.05, - "ImageVAEEncodingStage": 102.8, - "LatentPreparationStage": 0.21, - "TimestepPreparationStage": 14.55, - "DenoisingStage": 32205.84, - "DecodingStage": 36.48 + "InputValidationStage": 24.78, + "ImageEncodingStage": 714.55, + "ImageVAEEncodingStage": 51.33, + "LatentPreparationStage": 0.16, + "TimestepPreparationStage": 11.8, + "DenoisingStage": 31661.25, + "DecodingStage": 18.88 }, "denoise_step_ms": { - "0": 522.17, - "1": 649.49, - "2": 647.67, - "3": 644.17, - "4": 645.43, - "5": 646.28, - "6": 648.04, - "7": 645.36, - "8": 647.03, - "9": 647.65, - "10": 646.98, - "11": 646.02, - "12": 644.8, - "13": 646.24, - "14": 646.42, - "15": 645.12, - "16": 644.52, - "17": 645.91, - "18": 646.36, - "19": 646.21, - "20": 646.86, - "21": 646.68, - "22": 646.79, - "23": 645.93, - "24": 646.42, - "25": 646.85, - "26": 647.07, - "27": 648.51, - "28": 646.43, - "29": 645.89, - "30": 647.12, - "31": 646.36, - "32": 647.5, - "33": 646.62, - "34": 647.44, - "35": 646.88, - "36": 646.93, - "37": 645.37, - "38": 646.18, - "39": 647.18, - "40": 648.32, - "41": 646.36, - "42": 647.41, - "43": 646.95, - "44": 646.27, - "45": 646.3, - "46": 645.75, - "47": 646.96, - "48": 644.27, - "49": 644.37 + "0": 478.42, + "1": 639.36, + "2": 624.05, + "3": 641.68, + "4": 633.93, + "5": 629.7, + "6": 636.64, + "7": 632.61, + "8": 634.45, + "9": 639.15, + "10": 631.25, + "11": 637.01, + "12": 636.15, + "13": 633.45, + "14": 637.49, + "15": 636.16, + "16": 635.98, + "17": 633.24, + "18": 633.64, + "19": 635.55, + "20": 633.12, + "21": 634.69, + "22": 634.61, + "23": 633.84, + "24": 634.0, + "25": 637.6, + "26": 635.62, + "27": 636.5, + "28": 632.61, + "29": 635.94, + "30": 638.16, + "31": 634.96, + "32": 636.42, + "33": 640.92, + "34": 638.18, + "35": 637.46, + "36": 641.18, + "37": 634.96, + "38": 640.21, + "39": 641.08, + "40": 636.65, + "41": 640.47, + "42": 638.16, + "43": 636.12, + "44": 637.03, + "45": 640.74, + "46": 638.69, + "47": 643.17, + "48": 636.95, + "49": 632.72 }, - "expected_e2e_ms": 33638.49, - "expected_avg_denoise_ms": 644.0, - "expected_median_denoise_ms": 647.87, + "expected_e2e_ms": 33010.17, + "expected_avg_denoise_ms": 633.05, + "expected_median_denoise_ms": 636.16, "estimated_full_test_time_s": 153.6 }, "joyai_image_edit_ti2i": { "stages_ms": { - "InputValidationStage": 29.6, - "ImageEncodingStage": 740.87, - "ImageVAEEncodingStage": 70.47, + "InputValidationStage": 21.56, + "ImageEncodingStage": 612.88, + "ImageVAEEncodingStage": 67.95, "LatentPreparationStage": 0.17, - "TimestepPreparationStage": 20.66, - "DenoisingStage": 26894.18, + "TimestepPreparationStage": 15.19, + "DenoisingStage": 26686.38, "DecodingStage": 14.27 }, "denoise_step_ms": { "0": 432.34, "1": 673.28, - "2": 658.38, + "2": 657.39, "3": 677.9, - "4": 677.87, + "4": 671.52, "5": 665.09, - "6": 680.25, - "7": 678.54, - "8": 675.02, - "9": 683.38, - "10": 679.11, - "11": 674.55, - "12": 681.24, - "13": 680.79, - "14": 678.81, - "15": 680.94, - "16": 680.89, - "17": 678.07, - "18": 679.9, - "19": 682.67, - "20": 678.41, - "21": 679.92, - "22": 681.07, - "23": 679.93, - "24": 682.35, - "25": 680.8, - "26": 681.19, - "27": 682.05, - "28": 681.34, - "29": 680.8, - "30": 675.57, - "31": 679.21, - "32": 679.67, - "33": 675.05, - "34": 681.63, - "35": 678.62, - "36": 675.68, - "37": 678.58, - "38": 679.01, - "39": 677.64 + "6": 677.58, + "7": 672.65, + "8": 669.57, + "9": 674.81, + "10": 674.51, + "11": 670.56, + "12": 670.59, + "13": 670.77, + "14": 675.56, + "15": 673.46, + "16": 676.6, + "17": 673.97, + "18": 673.64, + "19": 675.71, + "20": 669.77, + "21": 672.28, + "22": 673.77, + "23": 673.28, + "24": 675.13, + "25": 672.96, + "26": 676.29, + "27": 676.38, + "28": 675.12, + "29": 671.67, + "30": 672.48, + "31": 671.58, + "32": 669.13, + "33": 670.99, + "34": 671.04, + "35": 673.41, + "36": 674.28, + "37": 673.07, + "38": 673.32, + "39": 667.02 }, - "expected_e2e_ms": 28350.2, - "expected_avg_denoise_ms": 672.19, - "expected_median_denoise_ms": 679.16, + "expected_e2e_ms": 27693.61, + "expected_avg_denoise_ms": 666.89, + "expected_median_denoise_ms": 673.29, "estimated_full_test_time_s": 117.6 }, "qwen_image_t2i_cache_dit_enabled": { "stages_ms": { "InputValidationStage": 0.04, - "TextEncodingStage": 540.0, + "TextEncodingStage": 231.2, "LatentPreparationStage": 0.16, "TimestepPreparationStage": 14.34, - "DenoisingStage": 4219.4, + "DenoisingStage": 4216.22, "DecodingStage": 8.11 }, "denoise_step_ms": { "0": 194.03, "1": 242.42, - "2": 244.66, + "2": 243.19, "3": 255.67, - "4": 62.91, - "5": 7.69, - "6": 226.06, + "4": 59.95, + "5": 6.59, + "6": 198.65, "7": 30.85, - "8": 7.48, - "9": 201.89, + "8": 6.38, + "9": 193.27, "10": 48.7, - "11": 7.51, - "12": 7.34, - "13": 200.01, + "11": 6.4, + "12": 6.64, + "13": 194.41, "14": 51.4, - "15": 7.51, - "16": 7.4, + "15": 6.56, + "16": 6.36, "17": 205.12, "18": 58.3, - "19": 7.53, - "20": 7.12, + "19": 6.55, + "20": 6.13, "21": 195.49, - "22": 56.76, - "23": 7.37, - "24": 7.28, - "25": 195.54, - "26": 56.73, - "27": 7.53, - "28": 6.92, + "22": 56.24, + "23": 6.36, + "24": 6.17, + "25": 194.51, + "26": 56.45, + "27": 6.37, + "28": 6.24, "29": 196.33, "30": 57.49, - "31": 7.33, - "32": 7.34, - "33": 201.88, + "31": 6.6, + "32": 6.24, + "33": 201.13, "34": 57.04, - "35": 7.23, - "36": 7.47, - "37": 198.44, + "35": 6.58, + "36": 6.45, + "37": 196.17, "38": 55.55, - "39": 7.07, + "39": 6.48, "40": 7.16, - "41": 196.03, + "41": 193.75, "42": 56.32, - "43": 6.54, - "44": 6.79, + "43": 6.43, + "44": 6.48, "45": 196.52, - "46": 58.6, - "47": 7.44, - "48": 212.42, + "46": 58.45, + "47": 6.55, + "48": 201.62, "49": 45.11 }, - "expected_e2e_ms": 4800.0, - "expected_avg_denoise_ms": 84.23, + "expected_e2e_ms": 4531.7, + "expected_avg_denoise_ms": 84.06, "expected_median_denoise_ms": 53.48, "estimated_full_test_time_s": 124.9 }, "wan2_1_t2v_1.3b_teacache_enabled": { "stages_ms": { - "TextEncodingStage": 1088.18, + "TextEncodingStage": 618.71, "DenoisingStage": 3926.63, - "InputValidationStage": 0.07, + "InputValidationStage": 0.05, "LatentPreparationStage": 0.14, "TimestepPreparationStage": 2.31, - "DecodingStage": 711.47 + "DecodingStage": 544.28 }, "denoise_step_ms": { "0": 80.49, - "1": 147.49, + "1": 125.15, "2": 144.75, "3": 144.41, "4": 143.82, - "5": 138.89, - "6": 36.55, + "5": 137.91, + "6": 33.18, "7": 102.12, - "8": 39.18, + "8": 32.52, "9": 104.07, "10": 37.91, - "11": 107.4, + "11": 105.12, "12": 2.52, - "13": 36.6, + "13": 31.09, "14": 106.91, - "15": 3.17, - "16": 41.13, + "15": 2.79, + "16": 30.23, "17": 95.25, "18": 2.67, - "19": 37.88, + "19": 29.76, "20": 110.46, "21": 2.66, - "22": 38.29, + "22": 29.96, "23": 108.86, "24": 2.7, - "25": 38.46, + "25": 32.5, "26": 108.62, - "27": 2.71, - "28": 41.23, + "27": 2.6, + "28": 29.72, "29": 107.12, - "30": 2.72, - "31": 36.52, + "30": 2.53, + "31": 30.15, "32": 112.05, "33": 2.5, - "34": 110.6, + "34": 104.27, "35": 38.24, - "36": 110.83, + "36": 102.6, "37": 37.63, - "38": 111.91, - "39": 37.14, - "40": 111.39, + "38": 104.91, + "39": 32.24, + "40": 106.69, "41": 38.64, - "42": 110.57, + "42": 106.66, "43": 38.09, - "44": 111.32, - "45": 148.48, - "46": 149.96, + "44": 109.63, + "45": 145.34, + "46": 148.45, "47": 143.63, "48": 148.49, - "49": 151.0 + "49": 150.48 }, - "expected_e2e_ms": 6012.96, + "expected_e2e_ms": 5161.84, "expected_avg_denoise_ms": 78.45, "expected_median_denoise_ms": 102.49, "estimated_full_test_time_s": 126.0 }, "wan2_1_t2v_1.3b": { "stages_ms": { - "DecodingStage": 632.0, - "InputValidationStage": 0.06, + "DecodingStage": 503.17, + "InputValidationStage": 0.03, "DenoisingStage": 7244.34, - "TextEncodingStage": 1083.56, - "LatentPreparationStage": 0.14, - "TimestepPreparationStage": 2.31 + "TextEncodingStage": 624.67, + "LatentPreparationStage": 0.09, + "TimestepPreparationStage": 2.28 }, "denoise_step_ms": { "0": 92.19, - "1": 137.78, + "1": 121.33, "2": 136.86, "3": 137.63, "4": 138.47, @@ -1134,7 +1134,7 @@ "48": 137.42, "49": 137.45 }, - "expected_e2e_ms": 9305.91, + "expected_e2e_ms": 8629.02, "expected_avg_denoise_ms": 144.79, "expected_median_denoise_ms": 145.57, "estimated_full_test_time_s": 129.3 @@ -1143,66 +1143,66 @@ "stages_ms": { "LatentPreparationStage": 0.29, "InputValidationStage": 0.05, - "DenoisingStage": 4934.97, - "DecodingStage": 631.04, + "DenoisingStage": 3735.53, + "DecodingStage": 404.36, "TimestepPreparationStage": 3.66, - "TextEncodingStage": 1450.0 + "TextEncodingStage": 652.39 }, "denoise_step_ms": { "0": 58.84, - "1": 65.73, - "2": 90.75, - "3": 91.42, - "4": 92.55, - "5": 92.37, - "6": 93.97, - "7": 92.8, - "8": 93.52, - "9": 90.21, - "10": 90.46, - "11": 91.8, - "12": 90.97, - "13": 91.41, - "14": 90.72, - "15": 90.27, - "16": 90.16, - "17": 90.02, - "18": 90.95, - "19": 89.52, - "20": 90.15, - "21": 89.55, - "22": 89.73, - "23": 89.52, - "24": 90.5, - "25": 91.02, - "26": 92.04, - "27": 91.13, - "28": 90.83, - "29": 91.43, - "30": 91.39, - "31": 91.71, - "32": 91.91, - "33": 92.12, - "34": 91.48, - "35": 91.61, - "36": 90.16, - "37": 90.03, - "38": 89.51, - "39": 92.34, - "40": 88.98, - "41": 89.95, - "42": 90.03, - "43": 89.52, - "44": 89.57, - "45": 89.83, - "46": 91.26, - "47": 91.85, - "48": 91.15, - "49": 91.1 + "1": 64.87, + "2": 66.46, + "3": 63.27, + "4": 67.23, + "5": 62.74, + "6": 60.87, + "7": 74.0, + "8": 75.26, + "9": 76.17, + "10": 73.24, + "11": 76.09, + "12": 87.22, + "13": 78.91, + "14": 76.16, + "15": 78.78, + "16": 75.38, + "17": 75.62, + "18": 75.29, + "19": 74.83, + "20": 74.68, + "21": 72.13, + "22": 73.38, + "23": 77.21, + "24": 77.8, + "25": 85.28, + "26": 74.18, + "27": 78.97, + "28": 77.76, + "29": 76.05, + "30": 73.97, + "31": 75.86, + "32": 75.52, + "33": 75.77, + "34": 74.81, + "35": 76.4, + "36": 76.53, + "37": 77.94, + "38": 78.81, + "39": 78.26, + "40": 78.24, + "41": 77.43, + "42": 76.87, + "43": 76.96, + "44": 76.15, + "45": 76.11, + "46": 75.26, + "47": 76.75, + "48": 79.17, + "49": 72.62 }, - "expected_e2e_ms": 6800.0, - "expected_avg_denoise_ms": 98.56, - "expected_median_denoise_ms": 100.05, + "expected_e2e_ms": 4956.47, + "expected_avg_denoise_ms": 74.54, + "expected_median_denoise_ms": 75.03, "estimated_full_test_time_s": 127.6 }, "turbo_wan2_1_t2v_1.3b": { @@ -1229,181 +1229,181 @@ "ltx_2_two_stage_t2v": { "stages_ms": { "InputValidationStage": 0.04, - "TextEncodingStage": 1570.42, + "TextEncodingStage": 775.78, "LTX2TextConnectorStage": 10.72, "LTX2HalveResolutionStage": 0.05, - "LTX2LoRASwitchStage": 34.59, + "LTX2LoRASwitchStage": 2.8, "LTX2SigmaPreparationStage": 0.17, "TimestepPreparationStage": 11.39, "LTX2AVLatentPreparationStage": 0.18, "LTX2ImageEncodingStage": 0.02, - "LTX2AVDenoisingStage": 4708.8, + "LTX2AVDenoisingStage": 4521.8, "LTX2UpsampleStage": 2.52, - "LTX2RefinementStage": 585.98, + "LTX2RefinementStage": 544.95, "LTX2AVDecodingStage": 206.92, "per_frame_generation": null }, "denoise_step_ms": { "0": 116.82, "1": 166.1, - "2": 121.53, - "3": 118.69, - "4": 121.32, + "2": 113.57, + "3": 116.56, + "4": 115.93, "5": 116.1, "6": 116.65, "7": 115.37, - "8": 113.66, + "8": 109.93, "9": 116.34, "10": 116.37, "11": 118.34, "12": 114.94, "13": 120.27, "14": 117.81, - "15": 118.44, + "15": 118.43, "16": 118.83, - "17": 115.56, + "17": 107.31, "18": 115.91, "19": 114.51, - "20": 116.32, - "21": 118.88, - "22": 115.94, - "23": 114.69, - "24": 115.36, - "25": 114.4, + "20": 106.1, + "21": 105.55, + "22": 111.25, + "23": 105.68, + "24": 105.36, + "25": 106.07, "26": 114.65, "27": 115.08, "28": 115.54, - "29": 115.14, - "30": 115.89, + "29": 114.29, + "30": 115.5, "31": 123.66, - "32": 120.16, - "33": 114.77, - "34": 114.4, - "35": 113.72, - "36": 114.14, - "37": 111.66, - "38": 113.4, - "39": 113.54, + "32": 104.07, + "33": 103.24, + "34": 104.41, + "35": 103.3, + "36": 102.05, + "37": 102.4, + "38": 102.77, + "39": 102.07, "40": 192.18, - "41": 193.62, - "42": 196.16 + "41": 171.85, + "42": 166.03 }, - "expected_e2e_ms": 7209.55, - "expected_avg_denoise_ms": 122.95, - "expected_median_denoise_ms": 115.94, + "expected_e2e_ms": 6313.7, + "expected_avg_denoise_ms": 118.58, + "expected_median_denoise_ms": 109.14, "estimated_full_test_time_s": 345.4 }, "wan2_2_ti2v_5b": { "stages_ms": { "InputValidationStage": 380.0, - "TextEncodingStage": 750.0, + "TextEncodingStage": 328.44, "LatentPreparationStage": 0.13, "TimestepPreparationStage": 2.32, - "DenoisingStage": 18728.92, - "DecodingStage": 1550.0 + "DenoisingStage": 15171.58, + "DecodingStage": 1527.14 }, "denoise_step_ms": { - "0": 225.06, - "1": 376.85, - "2": 366.99, - "3": 370.68, - "4": 361.07, - "5": 367.47, - "6": 367.77, - "7": 365.0, - "8": 370.13, - "9": 360.75, - "10": 367.07, - "11": 368.03, - "12": 365.97, - "13": 371.27, - "14": 373.7, - "15": 370.91, - "16": 361.13, - "17": 364.8, - "18": 370.17, - "19": 364.23, - "20": 368.85, - "21": 362.55, - "22": 364.73, - "23": 370.21, - "24": 363.45, - "25": 369.27, - "26": 363.87, - "27": 367.4, - "28": 369.43, - "29": 365.64, - "30": 369.69, - "31": 360.54, - "32": 367.64, - "33": 368.24, - "34": 365.06, - "35": 369.57, - "36": 365.15, - "37": 368.11, - "38": 364.87, - "39": 368.03, - "40": 367.75, - "41": 365.94, - "42": 369.07, - "43": 366.85, - "44": 367.81, - "45": 366.21, - "46": 370.75, - "47": 373.42, - "48": 373.94, - "49": 371.58 + "0": 183.83, + "1": 301.44, + "2": 300.81, + "3": 311.19, + "4": 305.2, + "5": 302.05, + "6": 307.57, + "7": 305.22, + "8": 299.57, + "9": 302.19, + "10": 306.41, + "11": 303.95, + "12": 304.63, + "13": 305.57, + "14": 304.51, + "15": 305.02, + "16": 305.67, + "17": 304.17, + "18": 305.43, + "19": 304.32, + "20": 305.91, + "21": 304.7, + "22": 302.28, + "23": 303.57, + "24": 305.08, + "25": 304.67, + "26": 304.03, + "27": 305.48, + "28": 306.68, + "29": 305.07, + "30": 304.83, + "31": 305.76, + "32": 304.22, + "33": 305.49, + "34": 306.05, + "35": 305.28, + "36": 306.09, + "37": 307.76, + "38": 305.25, + "39": 305.21, + "40": 306.17, + "41": 304.79, + "42": 305.64, + "43": 305.06, + "44": 306.74, + "45": 305.89, + "46": 304.28, + "47": 306.01, + "48": 304.94, + "49": 302.57 }, - "expected_e2e_ms": 18500.0, - "expected_avg_denoise_ms": 364.69, - "expected_median_denoise_ms": 367.69, + "expected_e2e_ms": 17597.35, + "expected_avg_denoise_ms": 302.53, + "expected_median_denoise_ms": 305.07, "estimated_full_test_time_s": 141.7 }, "qwen_image_edit_2509_ti2i": { "stages_ms": { - "ImageEncodingStage": 771.51, - "DenoisingStage": 38606.11, - "InputValidationStage": 123.03, - "LatentPreparationStage": 0.17, + "ImageEncodingStage": 590.9, + "DenoisingStage": 38451.75, + "InputValidationStage": 98.87, + "LatentPreparationStage": 0.15, "TimestepPreparationStage": 12.19, "DecodingStage": 14.52, - "ImageVAEEncodingStage": 183.82 + "ImageVAEEncodingStage": 91.27 }, "denoise_step_ms": { - "0": 725.94, - "1": 962.85, - "2": 965.28, - "3": 965.75, - "4": 966.83, - "5": 966.89, - "6": 965.49, - "7": 966.71, - "8": 970.34, - "9": 967.75, + "0": 717.34, + "1": 959.86, + "2": 964.82, + "3": 964.14, + "4": 964.39, + "5": 962.79, + "6": 959.88, + "7": 962.3, + "8": 959.08, + "9": 966.39, "10": 968.95, - "11": 967.74, - "12": 964.15, - "13": 967.79, - "14": 968.02, - "15": 971.74, - "16": 968.23, + "11": 966.02, + "12": 963.19, + "13": 965.61, + "14": 964.04, + "15": 964.3, + "16": 964.12, "17": 966.04, - "18": 971.09, - "19": 969.49, - "20": 968.51, - "21": 970.57, - "22": 969.32, - "23": 970.4, - "24": 972.18, - "25": 972.09, - "26": 970.16, - "27": 969.12, - "28": 971.7, + "18": 967.82, + "19": 968.01, + "20": 968.46, + "21": 968.58, + "22": 967.91, + "23": 967.14, + "24": 966.89, + "25": 967.93, + "26": 964.2, + "27": 968.56, + "28": 966.23, "29": 970.81, - "30": 970.75, + "30": 968.49, "31": 970.28, - "32": 970.21, - "33": 973.42, + "32": 968.94, + "33": 970.14, "34": 968.76, "35": 970.11, "36": 968.54, @@ -1411,331 +1411,331 @@ "38": 971.43, "39": 969.25 }, - "expected_e2e_ms": 40249.66, - "expected_avg_denoise_ms": 965.0, - "expected_median_denoise_ms": 971.69, + "expected_e2e_ms": 39483.04, + "expected_avg_denoise_ms": 961.13, + "expected_median_denoise_ms": 967.48, "estimated_full_test_time_s": 160.2 }, "qwen_image_layered_i2i": { "stages_ms": { - "QwenImageLayeredBeforeDenoisingStage": 132.71, + "QwenImageLayeredBeforeDenoisingStage": 800, "TimestepPreparationStage": 0.0, - "DenoisingStage": 33400.39, + "DenoisingStage": 33064.04, "DecodingStage": 142.89 }, "denoise_step_ms": { "0": 512.55, - "1": 675.96, - "2": 663.04, + "1": 671.77, + "2": 658.22, "3": 666.15, "4": 669.94, - "5": 668.69, + "5": 663.29, "6": 663.62, - "7": 672.62, - "8": 670.99, - "9": 672.66, - "10": 673.83, - "11": 671.75, - "12": 670.63, - "13": 672.97, - "14": 672.76, - "15": 675.31, - "16": 672.15, - "17": 671.69, - "18": 672.69, - "19": 670.83, - "20": 670.16, - "21": 671.72, + "7": 665.89, + "8": 663.92, + "9": 666.76, + "10": 662.27, + "11": 662.88, + "12": 667.29, + "13": 667.25, + "14": 668.65, + "15": 667.43, + "16": 667.93, + "17": 666.45, + "18": 670.42, + "19": 669.86, + "20": 666.7, + "21": 669.94, "22": 665.95, - "23": 670.78, - "24": 668.87, - "25": 672.74, - "26": 668.35, - "27": 672.91, + "23": 666.9, + "24": 668.24, + "25": 669.13, + "26": 667.71, + "27": 672.43, "28": 669.31, - "29": 669.8, - "30": 673.63, + "29": 660.48, + "30": 670.89, "31": 666.7, - "32": 670.55, + "32": 669.39, "33": 666.78, "34": 670.27, - "35": 675.04, + "35": 671.19, "36": 668.11, - "37": 671.81, - "38": 671.92, - "39": 669.62, + "37": 670.61, + "38": 670.22, + "39": 660.73, "40": 674.31, "41": 672.18, - "42": 670.9, + "42": 670.49, "43": 669.57, - "44": 674.92, + "44": 674.7, "45": 673.28, - "46": 674.95, - "47": 673.68, - "48": 674.14, - "49": 670.44 + "46": 672.62, + "47": 672.52, + "48": 673.85, + "49": 666.93 }, - "expected_e2e_ms": 33679.2, - "expected_avg_denoise_ms": 667.88, - "expected_median_denoise_ms": 671.34, + "expected_e2e_ms": 33451.98, + "expected_avg_denoise_ms": 661.05, + "expected_median_denoise_ms": 664.26, "estimated_full_test_time_s": 161.5 }, "fastwan2_2_ti2v_5b": { "stages_ms": { "InputValidationStage": 400.0, - "TextEncodingStage": 685.05, - "TimestepPreparationStage": 36.07, + "TextEncodingStage": 329.33, + "TimestepPreparationStage": 30.08, "LatentPreparationStage": 0.12, - "DmdDenoisingStage": 401.85, - "DecodingStage": 1524.82 + "DmdDenoisingStage": 270.39, + "DecodingStage": 1462.06 }, "denoise_step_ms": { "0": 35.0, - "1": 115.0, - "2": 130.0 + "1": 30.01, + "2": 114.63 }, - "expected_e2e_ms": 3150.0, - "expected_avg_denoise_ms": 95.0, - "expected_median_denoise_ms": 120.0, + "expected_e2e_ms": 2680.2, + "expected_avg_denoise_ms": 88.05, + "expected_median_denoise_ms": 114.63, "estimated_full_test_time_s": 125.2 }, "fast_hunyuan_video": { "stages_ms": { - "InputValidationStage": 0.06, - "TextEncodingStage": 300.0, + "InputValidationStage": 0.04, + "TextEncodingStage": 254.74, "TimestepPreparationStage": 28.98, "LatentPreparationStage": 0.13, - "DenoisingStage": 5898.72, - "DecodingStage": 4736.19, + "DenoisingStage": 5812.04, + "DecodingStage": 2888.31, "per_frame_generation": null }, "denoise_step_ms": { "0": 286.89, - "1": 1115.43, - "2": 1118.06, - "3": 1124.53, - "4": 1130.91, - "5": 1119.61 + "1": 284.01, + "2": 1090.2, + "3": 1094.58, + "4": 1087.51, + "5": 1090.66 }, - "expected_e2e_ms": 11816.27, - "expected_avg_denoise_ms": 982.57, - "expected_median_denoise_ms": 1118.83, + "expected_e2e_ms": 9761.46, + "expected_avg_denoise_ms": 967.68, + "expected_median_denoise_ms": 1090.43, "estimated_full_test_time_s": 77.0 }, "wan2_2_i2v_a14b_2gpu": { "stages_ms": { "InputValidationStage": 18.1, - "ImageVAEEncodingStage": 1719.86, - "DecodingStage": 1990.63, + "ImageVAEEncodingStage": 1630.35, + "DecodingStage": 1114.63, "LatentPreparationStage": 0.31, - "TextEncodingStage": 1145.1, + "TextEncodingStage": 626.49, "TimestepPreparationStage": 4.42, - "DenoisingStage": 69317.11 + "DenoisingStage": 68736.39 }, "denoise_step_ms": { "0": 519.5, - "1": 2058.07, - "2": 1770.74, - "3": 1716.11, + "1": 612.04, + "2": 1737.52, + "3": 1705.6, "4": 1711.89, - "5": 1720.71, - "6": 1726.19, - "7": 1719.55, - "8": 1722.27, - "9": 1723.75, - "10": 1729.27, - "11": 1726.47, - "12": 1730.2, - "13": 1729.31, - "14": 1731.41, - "15": 2791.26, - "16": 1795.33, - "17": 1731.09, - "18": 1720.22, - "19": 1722.62, - "20": 1728.4, - "21": 1724.47, - "22": 1723.9, - "23": 1726.33, - "24": 1726.05, - "25": 1727.68, - "26": 1724.88, - "27": 1728.33, - "28": 1729.43, - "29": 1721.09, - "30": 1723.38, - "31": 1720.92, - "32": 1726.09, - "33": 1724.37, - "34": 1724.11, - "35": 1722.94, - "36": 1722.8, - "37": 1721.07, - "38": 1719.73, - "39": 1700.15 + "5": 1714.28, + "6": 1711.38, + "7": 1710.32, + "8": 1714.7, + "9": 1715.46, + "10": 1713.3, + "11": 1714.58, + "12": 1716.05, + "13": 1718.95, + "14": 1720.84, + "15": 1743.12, + "16": 1715.18, + "17": 1718.87, + "18": 1712.45, + "19": 1712.11, + "20": 1711.92, + "21": 1715.97, + "22": 1712.68, + "23": 1710.74, + "24": 1715.36, + "25": 1715.06, + "26": 1712.65, + "27": 1710.31, + "28": 1714.64, + "29": 1714.87, + "30": 1714.28, + "31": 1713.73, + "32": 1714.59, + "33": 1711.81, + "34": 1716.93, + "35": 1716.84, + "36": 1715.25, + "37": 1715.56, + "38": 1713.6, + "39": 1697.24 }, - "expected_e2e_ms": 76555.66, - "expected_avg_denoise_ms": 1731.55, - "expected_median_denoise_ms": 1724.42, + "expected_e2e_ms": 73811.61, + "expected_avg_denoise_ms": 1689.73, + "expected_median_denoise_ms": 1714.58, "estimated_full_test_time_s": 264.1 }, "wan2_1_i2v_14b_480P_2gpu": { "stages_ms": { "InputValidationStage": 7.88, - "LatentPreparationStage": 0.21, - "ImageEncodingStage": 1552.03, - "DenoisingStage": 118408.29, - "DecodingStage": 748.39, - "TimestepPreparationStage": 2.21, - "TextEncodingStage": 1001.21, - "ImageVAEEncodingStage": 498.16 + "LatentPreparationStage": 0.19, + "ImageEncodingStage": 30.99, + "DenoisingStage": 33259.34, + "DecodingStage": 570.09, + "TimestepPreparationStage": 2.14, + "TextEncodingStage": 400.75, + "ImageVAEEncodingStage": 451.69 }, "denoise_step_ms": { - "0": 1799.07, - "1": 2345.85, - "2": 2362.11, - "3": 2368.42, - "4": 2379.09, - "5": 2382.94, - "6": 2383.58, - "7": 2383.3, - "8": 2372.44, - "9": 2370.42, - "10": 2367.9, - "11": 2365.91, - "12": 2365.92, - "13": 2363.78, - "14": 2361.17, - "15": 2359.55, - "16": 2360.3, - "17": 2358.31, - "18": 2358.78, - "19": 2357.74, - "20": 2356.39, - "21": 2355.8, - "22": 2354.59, - "23": 2354.24, - "24": 2356.86, - "25": 2354.13, - "26": 2350.14, - "27": 2351.26, - "28": 2349.73, - "29": 2351.05, - "30": 2348.73, - "31": 2350.26, - "32": 2352.1, - "33": 2349.96, - "34": 2350.83, - "35": 2350.85, - "36": 2350.63, - "37": 2352.8, - "38": 2349.64, - "39": 2351.67, - "40": 2351.21, - "41": 2350.95, - "42": 2350.02, - "43": 2348.38, - "44": 2348.84, - "45": 2347.04, - "46": 2350.02, - "47": 2347.09, - "48": 2345.31, - "49": 2344.33 + "0": 511.75, + "1": 684.53, + "2": 655.82, + "3": 667.35, + "4": 669.71, + "5": 664.06, + "6": 667.2, + "7": 669.88, + "8": 662.88, + "9": 665.68, + "10": 665.83, + "11": 665.85, + "12": 663.76, + "13": 666.65, + "14": 663.7, + "15": 666.51, + "16": 667.55, + "17": 666.8, + "18": 670.57, + "19": 670.01, + "20": 667.19, + "21": 667.71, + "22": 666.87, + "23": 668.3, + "24": 668.23, + "25": 668.41, + "26": 668.06, + "27": 665.42, + "28": 671.21, + "29": 669.95, + "30": 667.59, + "31": 670.01, + "32": 667.18, + "33": 670.25, + "34": 669.73, + "35": 670.18, + "36": 671.23, + "37": 671.36, + "38": 670.11, + "39": 668.64, + "40": 668.9, + "41": 670.84, + "42": 667.72, + "43": 670.59, + "44": 670.31, + "45": 667.49, + "46": 668.46, + "47": 670.64, + "48": 666.37, + "49": 662.19 }, - "expected_e2e_ms": 122993.8, - "expected_avg_denoise_ms": 2368.05, - "expected_median_denoise_ms": 2379.34, + "expected_e2e_ms": 34741.5, + "expected_avg_denoise_ms": 665.03, + "expected_median_denoise_ms": 667.89, "estimated_full_test_time_s": 243.0 }, "wan2_1_i2v_14b_720P_2gpu": { "stages_ms": { "InputValidationStage": 11.58, - "TextEncodingStage": 998.23, + "TextEncodingStage": 390.5, "DecodingStage": 1357.99, "TimestepPreparationStage": 2.27, - "ImageEncodingStage": 998.9, + "ImageEncodingStage": 22.34, "LatentPreparationStage": 0.17, - "DenoisingStage": 120583.76, - "ImageVAEEncodingStage": 1115.82 + "DenoisingStage": 98693.53, + "ImageVAEEncodingStage": 938.33 }, "denoise_step_ms": { - "0": 1823.9, - "1": 2415.99, - "2": 2412.06, - "3": 2425.42, - "4": 2425.92, - "5": 2409.47, - "6": 2418.57, - "7": 2428.36, - "8": 2410.78, - "9": 2415.61, - "10": 2428.53, - "11": 2414.33, - "12": 2412.63, - "13": 2429.5, - "14": 2417.51, - "15": 2409.05, - "16": 2427.37, - "17": 2419.22, - "18": 2409.34, - "19": 2424.76, - "20": 2428.12, - "21": 2406.81, - "22": 2417.63, - "23": 2426.56, - "24": 2407.51, - "25": 2415.62, - "26": 2427.82, - "27": 2412.96, - "28": 2411.93, - "29": 2428.2, - "30": 2415.34, - "31": 2408.42, - "32": 2424.59, - "33": 2422.08, - "34": 2408.36, - "35": 2424.37, - "36": 2429.94, - "37": 2424.1, - "38": 2428.31, - "39": 2428.07, - "40": 2428.73, - "41": 2431.32, - "42": 2426.16, - "43": 2425.57, - "44": 2425.57, - "45": 2428.39, - "46": 2425.79, - "47": 2416.71, - "48": 2426.8, - "49": 2415.26 + "0": 1520.16, + "1": 2022.96, + "2": 1995.39, + "3": 1991.55, + "4": 1991.27, + "5": 1992.33, + "6": 1995.09, + "7": 1991.43, + "8": 1996.42, + "9": 1995.96, + "10": 1979.55, + "11": 1996.81, + "12": 1995.44, + "13": 1997.31, + "14": 1998.39, + "15": 2001.51, + "16": 1996.36, + "17": 2004.52, + "18": 1998.34, + "19": 1996.5, + "20": 1989.38, + "21": 1999.1, + "22": 2001.81, + "23": 2003.79, + "24": 2000.16, + "25": 2001.36, + "26": 1997.9, + "27": 1998.86, + "28": 2001.61, + "29": 1984.19, + "30": 2001.64, + "31": 2001.72, + "32": 1999.12, + "33": 1999.24, + "34": 1999.12, + "35": 2000.89, + "36": 1997.08, + "37": 1997.03, + "38": 1998.96, + "39": 1986.55, + "40": 2000.45, + "41": 2000.86, + "42": 1997.49, + "43": 2000.21, + "44": 1995.28, + "45": 1998.78, + "46": 1997.67, + "47": 1997.41, + "48": 1990.79, + "49": 1970.68 }, - "expected_e2e_ms": 126670.57, - "expected_avg_denoise_ms": 2411.52, - "expected_median_denoise_ms": 2424.24, + "expected_e2e_ms": 101519.43, + "expected_avg_denoise_ms": 1973.69, + "expected_median_denoise_ms": 1982.25, "estimated_full_test_time_s": 248.9 }, "wan2_2_t2v_a14b_2gpu": { "stages_ms": { "InputValidationStage": 0.05, - "TextEncodingStage": 800.0, + "TextEncodingStage": 314.11, "LatentPreparationStage": 0.21, "TimestepPreparationStage": 1.89, - "DenoisingStage": 82060.15, - "DecodingStage": 1153.08 + "DenoisingStage": 81747.14, + "DecodingStage": 784.99 }, "denoise_step_ms": { "0": 1453.25, "1": 2056.66, - "2": 2033.25, + "2": 2027.99, "3": 2012.96, - "4": 2020.89, + "4": 2020.39, "5": 2019.44, - "6": 2082.17, - "7": 2156.95, - "8": 2157.94, - "9": 2143.51, - "10": 2139.77, + "6": 2056.54, + "7": 2065.04, + "8": 2020.14, + "9": 2052.19, + "10": 2020.63, "11": 2075.95, "12": 2020.56, "13": 2020.28, @@ -1751,104 +1751,104 @@ "23": 2019.69, "24": 2019.61, "25": 2019.24, - "26": 2823.79, + "26": 2084.71, "27": 2022.36, "28": 2031.39, "29": 2048.32, - "30": 2045.05, - "31": 2051.02, - "32": 2046.83, - "33": 2046.99, - "34": 2045.64, - "35": 2047.25, - "36": 2048.41, - "37": 2044.96, + "30": 2020.36, + "31": 2033.97, + "32": 2036.4, + "33": 2039.62, + "34": 2045.06, + "35": 2026.33, + "36": 2022.31, + "37": 2026.32, "38": 2020.56, - "39": 2057.8 + "39": 2014.28 }, - "expected_e2e_ms": 85084.03, - "expected_avg_denoise_ms": 2050.21, + "expected_e2e_ms": 83297.78, + "expected_avg_denoise_ms": 2029.88, "expected_median_denoise_ms": 2026.87, "estimated_full_test_time_s": 204.3 }, "wan2_1_t2v_14b_2gpu": { "stages_ms": { - "TextEncodingStage": 1450.0, + "TextEncodingStage": 648.05, "DecodingStage": 637.43, "TimestepPreparationStage": 3.4, "InputValidationStage": 0.05, "LatentPreparationStage": 0.27, - "DenoisingStage": 49954.12 + "DenoisingStage": 23262.32 }, "denoise_step_ms": { "0": 307.85, - "1": 992.37, - "2": 996.76, - "3": 998.41, - "4": 1000.11, - "5": 998.95, - "6": 998.69, - "7": 998.63, - "8": 998.84, - "9": 1002.57, - "10": 997.34, - "11": 996.27, - "12": 997.68, - "13": 997.53, - "14": 998.96, - "15": 997.3, - "16": 1000.01, - "17": 996.27, - "18": 998.87, - "19": 997.53, - "20": 995.63, - "21": 998.22, - "22": 997.49, - "23": 998.05, - "24": 999.49, - "25": 999.94, - "26": 998.89, - "27": 996.19, - "28": 997.62, - "29": 997.51, - "30": 997.65, - "31": 997.51, - "32": 997.47, - "33": 995.01, - "34": 996.29, - "35": 997.55, - "36": 996.31, - "37": 997.5, - "38": 995.01, - "39": 996.42, - "40": 997.68, - "41": 996.08, - "42": 994.98, - "43": 997.63, - "44": 995.02, - "45": 995.25, - "46": 994.79, - "47": 996.51, - "48": 996.08, - "49": 996.27 + "1": 190.41, + "2": 490.26, + "3": 468.77, + "4": 479.98, + "5": 470.02, + "6": 481.03, + "7": 474.6, + "8": 480.52, + "9": 475.21, + "10": 466.27, + "11": 475.87, + "12": 476.53, + "13": 477.88, + "14": 476.46, + "15": 478.46, + "16": 475.68, + "17": 480.34, + "18": 473.5, + "19": 481.12, + "20": 469.81, + "21": 480.89, + "22": 473.67, + "23": 478.55, + "24": 475.17, + "25": 478.69, + "26": 476.86, + "27": 476.79, + "28": 478.15, + "29": 469.87, + "30": 478.06, + "31": 478.14, + "32": 477.48, + "33": 478.13, + "34": 476.76, + "35": 476.79, + "36": 478.36, + "37": 476.93, + "38": 477.99, + "39": 469.25, + "40": 477.28, + "41": 478.29, + "42": 476.71, + "43": 477.91, + "44": 479.14, + "45": 477.02, + "46": 477.17, + "47": 478.45, + "48": 476.97, + "49": 464.05 }, - "expected_e2e_ms": 53244.4, - "expected_avg_denoise_ms": 998.96, - "expected_median_denoise_ms": 1010.02, + "expected_e2e_ms": 24749.27, + "expected_avg_denoise_ms": 464.73, + "expected_median_denoise_ms": 470.88, "estimated_full_test_time_s": 173.2 }, "wan2_2_t2v_a14b_lora_2gpu": { "stages_ms": { - "InputValidationStage": 0.06, - "TextEncodingStage": 1500.0, - "LatentPreparationStage": 0.15, - "TimestepPreparationStage": 4.13, + "InputValidationStage": 0.03, + "TextEncodingStage": 616.22, + "LatentPreparationStage": 0.13, + "TimestepPreparationStage": 2.72, "DenoisingStage": 57638.76, - "DecodingStage": 1659.71 + "DecodingStage": 823.49 }, "denoise_step_ms": { "0": 504.86, - "1": 1482.77, + "1": 1188.64, "2": 1422.03, "3": 1404.94, "4": 1421.26, @@ -1873,8 +1873,8 @@ "23": 1419.4, "24": 1417.73, "25": 1417.12, - "26": 2656.31, - "27": 1933.46, + "26": 1619.69, + "27": 1598.62, "28": 1426.79, "29": 1423.25, "30": 1411.6, @@ -1896,7 +1896,7 @@ "wan2_1_t2v_1_3b_lora_1gpu": { "stages_ms": { "InputValidationStage": 0.05, - "TextEncodingStage": 679.6, + "TextEncodingStage": 340.58, "LatentPreparationStage": 0.16, "TimestepPreparationStage": 3.26, "DenoisingStage": 7328.48, @@ -1904,15 +1904,15 @@ }, "denoise_step_ms": { "0": 90.0, - "1": 141.31, + "1": 119.02, "2": 146.15, - "3": 145.23, + "3": 144.15, "4": 144.57, "5": 144.41, "6": 160.18, "7": 148.95, "8": 144.97, - "9": 146.95, + "9": 145.9, "10": 144.53, "11": 145.21, "12": 146.25, @@ -1926,35 +1926,35 @@ "20": 151.05, "21": 146.67, "22": 147.05, - "23": 147.91, - "24": 148.1, + "23": 145.42, + "24": 147.84, "25": 147.86, "26": 149.47, "27": 150.5, - "28": 149.59, + "28": 148.85, "29": 145.76, - "30": 147.28, - "31": 147.92, - "32": 148.82, + "30": 144.58, + "31": 147.04, + "32": 148.56, "33": 149.74, "34": 148.97, "35": 147.51, - "36": 147.15, - "37": 148.49, - "38": 148.15, + "36": 147.01, + "37": 146.34, + "38": 147.05, "39": 148.62, "40": 148.49, "41": 149.09, "42": 146.27, - "43": 147.88, + "43": 145.72, "44": 145.54, - "45": 149.11, - "46": 149.56, - "47": 149.23, + "45": 148.21, + "46": 148.05, + "47": 148.89, "48": 146.76, - "49": 146.49 + "49": 145.83 }, - "expected_e2e_ms": 8579.79, + "expected_e2e_ms": 8324.93, "expected_avg_denoise_ms": 146.48, "expected_median_denoise_ms": 147.68, "estimated_full_test_time_s": 129.6 @@ -1962,84 +1962,84 @@ "wan2_1_i2v_14b_lora_2gpu": { "stages_ms": { "InputValidationStage": 12.06, - "TextEncodingStage": 1002.37, + "TextEncodingStage": 341.83, "DecodingStage": 1329.75, "TimestepPreparationStage": 2.14, - "ImageEncodingStage": 1149.06, - "LatentPreparationStage": 0.19, - "DenoisingStage": 121336.18, - "ImageVAEEncodingStage": 981.21 + "ImageEncodingStage": 33.35, + "LatentPreparationStage": 0.17, + "DenoisingStage": 98821.24, + "ImageVAEEncodingStage": 936.91 }, "denoise_step_ms": { - "0": 1849.25, - "1": 2398.42, - "2": 2404.04, - "3": 2402.96, - "4": 2408.18, - "5": 2406.33, - "6": 2405.18, - "7": 2402.38, - "8": 2405.01, - "9": 2400.34, - "10": 2404.48, - "11": 2401.93, - "12": 2403.85, - "13": 2402.69, - "14": 2406.45, - "15": 2404.48, - "16": 2401.97, - "17": 2404.2, - "18": 2407.47, - "19": 2402.91, - "20": 2403.42, - "21": 2403.94, - "22": 2403.89, - "23": 2403.06, - "24": 2405.37, - "25": 2407.47, - "26": 2404.17, - "27": 2403.18, - "28": 2403.49, - "29": 2402.85, - "30": 2403.01, - "31": 2403.22, - "32": 2403.09, - "33": 2403.72, - "34": 2402.49, - "35": 2404.09, - "36": 2403.63, - "37": 2402.67, - "38": 2402.19, - "39": 2403.69, - "40": 2403.46, - "41": 2403.39, - "42": 2402.44, - "43": 2403.48, - "44": 2409.01, - "45": 2407.25, - "46": 2407.37, - "47": 2404.01, - "48": 2401.44, - "49": 2405.4 + "0": 1536.32, + "1": 2013.2, + "2": 1986.65, + "3": 1975.67, + "4": 1976.62, + "5": 1978.03, + "6": 1977.88, + "7": 1979.89, + "8": 1981.97, + "9": 1978.5, + "10": 1982.51, + "11": 1982.54, + "12": 1984.15, + "13": 1987.31, + "14": 1983.76, + "15": 1988.0, + "16": 1988.23, + "17": 1982.14, + "18": 1983.82, + "19": 1986.63, + "20": 1987.46, + "21": 1984.61, + "22": 1985.56, + "23": 1987.15, + "24": 1986.5, + "25": 1986.78, + "26": 1989.74, + "27": 1987.21, + "28": 1986.19, + "29": 1988.72, + "30": 1991.59, + "31": 1985.31, + "32": 1985.79, + "33": 1986.29, + "34": 1989.01, + "35": 1988.83, + "36": 1986.41, + "37": 1990.63, + "38": 1989.91, + "39": 1987.18, + "40": 1986.08, + "41": 1991.05, + "42": 1983.31, + "43": 1984.62, + "44": 1984.96, + "45": 1984.37, + "46": 1981.33, + "47": 1980.99, + "48": 1982.94, + "49": 1967.75 }, - "expected_e2e_ms": 127413.82, - "expected_avg_denoise_ms": 2426.52, - "expected_median_denoise_ms": 2437.55, + "expected_e2e_ms": 101924.55, + "expected_avg_denoise_ms": 1976.24, + "expected_median_denoise_ms": 1985.67, "estimated_full_test_time_s": 248.2 }, "flux_2_image_t2i_2_gpus": { "stages_ms": { - "InputValidationStage": 0.28, - "TextEncodingStage": 820.0, + "InputValidationStage": 0.05, + "TextEncodingStage": 747.5, "ImageVAEEncodingStage": 0.01, "LatentPreparationStage": 0.83, - "TimestepPreparationStage": 46.57, - "DenoisingStage": 13642.48, - "DecodingStage": 35.06 + "TimestepPreparationStage": 30.12, + "DenoisingStage": 13360.47, + "DecodingStage": 19.32 }, "denoise_step_ms": { "0": 66.76, - "1": 226.11, + "1": 71.94, "2": 226.39, "3": 226.37, "4": 226.46, @@ -2089,91 +2089,91 @@ "48": 227.69, "49": 227.17 }, - "expected_e2e_ms": 15433.36, - "expected_avg_denoise_ms": 268.04, - "expected_median_denoise_ms": 272.27, + "expected_e2e_ms": 15050.94, + "expected_avg_denoise_ms": 262.66, + "expected_median_denoise_ms": 266.23, "estimated_full_test_time_s": 135.3 }, "qwen_image_edit_2511_ti2i": { "stages_ms": { - "DecodingStage": 19.39, - "InputValidationStage": 54.97, - "DenoisingStage": 22253.55, - "ImageEncodingStage": 733.67, + "DecodingStage": 18.69, + "InputValidationStage": 48.98, + "DenoisingStage": 22129.32, + "ImageEncodingStage": 562.02, "LatentPreparationStage": 0.13, "TimestepPreparationStage": 12.36, - "ImageVAEEncodingStage": 84.65 + "ImageVAEEncodingStage": 47.89 }, "denoise_step_ms": { - "0": 416.33, - "1": 566.53, + "0": 415.55, + "1": 558.04, "2": 548.65, - "3": 566.13, - "4": 554.12, + "3": 560.17, + "4": 550.15, "5": 554.87, - "6": 555.5, + "6": 548.51, "7": 554.21, - "8": 558.77, + "8": 551.19, "9": 557.15, - "10": 560.32, + "10": 553.7, "11": 556.73, - "12": 558.26, - "13": 558.58, - "14": 558.03, - "15": 557.85, + "12": 554.35, + "13": 555.94, + "14": 556.44, + "15": 554.15, "16": 554.86, - "17": 558.62, - "18": 560.16, - "19": 560.23, - "20": 559.69, - "21": 559.95, - "22": 557.1, - "23": 560.04, - "24": 558.76, - "25": 559.81, - "26": 559.67, - "27": 558.74, - "28": 559.0, - "29": 559.09, + "17": 555.14, + "18": 558.16, + "19": 554.71, + "20": 557.93, + "21": 556.35, + "22": 555.88, + "23": 558.05, + "24": 556.66, + "25": 557.84, + "26": 556.77, + "27": 558.22, + "28": 555.72, + "29": 558.04, "30": 555.66, - "31": 559.22, - "32": 558.76, - "33": 560.83, + "31": 559.13, + "32": 555.84, + "33": 559.4, "34": 557.41, - "35": 560.1, + "35": 559.05, "36": 558.7, - "37": 560.89, - "38": 557.72, - "39": 559.24 + "37": 556.78, + "38": 556.8, + "39": 554.2 }, - "expected_e2e_ms": 23405.08, - "expected_avg_denoise_ms": 556.18, - "expected_median_denoise_ms": 560.34, + "expected_e2e_ms": 23143.72, + "expected_avg_denoise_ms": 552.98, + "expected_median_denoise_ms": 556.77, "estimated_full_test_time_s": 143.7 }, "fsdp-inference": { "stages_ms": { - "InputValidationStage": 0.06, + "InputValidationStage": 0.04, "LatentPreparationStage": 0.16, - "TextEncodingStage": 280.0, - "TimestepPreparationStage": 57.19, + "TextEncodingStage": 257.5, + "TimestepPreparationStage": 38.44, "DecodingStage": 16.88, - "DenoisingStage": 2422.53 + "DenoisingStage": 2374.14 }, "denoise_step_ms": { "0": 259.26, - "1": 284.25, - "2": 283.74, - "3": 270.48, - "4": 278.55, - "5": 271.58, - "6": 270.89, - "7": 277.75, - "8": 270.1 + "1": 164.46, + "2": 263.36, + "3": 257.61, + "4": 271.46, + "5": 266.71, + "6": 232.55, + "7": 270.08, + "8": 259.96 }, "expected_e2e_ms": 2775.88, - "expected_avg_denoise_ms": 268.55, - "expected_median_denoise_ms": 268.51, + "expected_avg_denoise_ms": 263.19, + "expected_median_denoise_ms": 261.85, "estimated_full_test_time_s": 122.7 }, "hunyuan3d_shape_gen": { @@ -2246,7 +2246,7 @@ "wan2_1_t2v_1.3b_frame_interp_2x": { "stages_ms": { "InputValidationStage": 0.05, - "TextEncodingStage": 677.98, + "TextEncodingStage": 612.62, "LatentPreparationStage": 0.2, "TimestepPreparationStage": 2.72, "DenoisingStage": 7336.03, @@ -2254,7 +2254,7 @@ }, "denoise_step_ms": { "0": 88.11, - "1": 143.49, + "1": 122.29, "2": 146.56, "3": 145.57, "4": 145.12, @@ -2266,33 +2266,33 @@ "10": 145.28, "11": 144.48, "12": 150.64, - "13": 154.2, + "13": 153.84, "14": 147.59, "15": 146.91, - "16": 148.0, + "16": 147.58, "17": 144.84, - "18": 147.03, - "19": 151.16, - "20": 150.71, + "18": 146.74, + "19": 150.18, + "20": 150.08, "21": 146.22, "22": 146.88, "23": 145.99, "24": 145.55, "25": 148.05, "26": 149.89, - "27": 149.75, + "27": 148.3, "28": 148.18, - "29": 147.27, + "29": 146.65, "30": 147.5, "31": 145.22, "32": 149.73, - "33": 151.0, + "33": 150.95, "34": 146.49, "35": 147.33, - "36": 148.9, + "36": 146.74, "37": 145.63, "38": 147.07, - "39": 149.91, + "39": 148.72, "40": 150.98, "41": 146.97, "42": 148.43, @@ -2301,18 +2301,18 @@ "45": 149.13, "46": 149.84, "47": 150.0, - "48": 146.38, - "49": 147.81 + "48": 145.6, + "49": 146.37 }, - "expected_e2e_ms": 8605.48, + "expected_e2e_ms": 8543.5, "expected_avg_denoise_ms": 146.61, "expected_median_denoise_ms": 147.3, "estimated_full_test_time_s": 129.3 }, "flux_2_image_t2i_upscaling_4x": { "stages_ms": { - "TextEncodingStage": 430.0, - "DenoisingStage": 23822.05, + "TextEncodingStage": 374.99, + "DenoisingStage": 23676.19, "InputValidationStage": 0.06, "LatentPreparationStage": 1.2, "TimestepPreparationStage": 18.66, @@ -2321,75 +2321,75 @@ }, "denoise_step_ms": { "0": 66.32, - "1": 453.87, + "1": 395.03, "2": 488.73, - "3": 467.7, - "4": 476.52, - "5": 475.02, + "3": 460.96, + "4": 474.42, + "5": 471.59, "6": 469.96, "7": 472.41, "8": 468.92, - "9": 477.5, + "9": 475.33, "10": 465.3, "11": 477.98, - "12": 469.2, - "13": 478.9, - "14": 468.38, - "15": 479.7, + "12": 466.21, + "13": 478.19, + "14": 467.91, + "15": 477.69, "16": 468.21, - "17": 479.16, - "18": 472.24, - "19": 480.48, - "20": 472.46, - "21": 478.22, - "22": 477.53, - "23": 476.9, - "24": 477.44, - "25": 475.33, - "26": 478.39, - "27": 476.84, - "28": 479.72, - "29": 476.64, - "30": 476.38, - "31": 476.48, - "32": 477.84, - "33": 475.79, - "34": 479.66, - "35": 478.66, - "36": 479.74, - "37": 479.39, - "38": 480.43, - "39": 477.42, - "40": 478.53, - "41": 477.29, - "42": 477.02, - "43": 477.34, - "44": 477.42, - "45": 477.18, - "46": 476.49, - "47": 477.06, + "17": 477.27, + "18": 470.51, + "19": 473.63, + "20": 472.38, + "21": 472.47, + "22": 475.89, + "23": 471.61, + "24": 474.66, + "25": 471.33, + "26": 476.31, + "27": 473.64, + "28": 474.4, + "29": 472.35, + "30": 474.65, + "31": 471.61, + "32": 476.36, + "33": 468.74, + "34": 473.22, + "35": 471.23, + "36": 476.13, + "37": 472.25, + "38": 473.52, + "39": 473.61, + "40": 474.92, + "41": 474.98, + "42": 473.82, + "43": 473.32, + "44": 474.41, + "45": 473.61, + "46": 473.71, + "47": 475.03, "48": 473.97, - "49": 476.92 + "49": 473.92 }, - "expected_e2e_ms": 24808.01, - "expected_avg_denoise_ms": 468.18, - "expected_median_denoise_ms": 477.12, + "expected_e2e_ms": 24551.25, + "expected_avg_denoise_ms": 465.17, + "expected_median_denoise_ms": 473.63, "estimated_full_test_time_s": 145.1 }, "wan2_1_t2v_1.3b_upscaling_4x": { "stages_ms": { - "DecodingStage": 694.93, - "InputValidationStage": 0.06, + "DecodingStage": 566.48, + "InputValidationStage": 0.04, "DenoisingStage": 7238.94, - "TextEncodingStage": 1077.43, + "TextEncodingStage": 623.45, "LatentPreparationStage": 0.14, "TimestepPreparationStage": 2.35 }, "denoise_step_ms": { "0": 85.84, - "1": 143.52, + "1": 104.79, "2": 143.85, - "3": 145.76, + "3": 144.68, "4": 145.57, "5": 144.43, "6": 145.13, @@ -2402,21 +2402,21 @@ "13": 148.34, "14": 144.98, "15": 143.98, - "16": 146.26, + "16": 145.08, "17": 143.22, "18": 145.69, "19": 144.76, "20": 149.62, "21": 139.68, "22": 144.04, - "23": 145.62, + "23": 144.42, "24": 144.34, "25": 145.08, "26": 144.94, "27": 145.77, "28": 142.06, "29": 146.45, - "30": 146.16, + "30": 145.53, "31": 143.51, "32": 144.69, "33": 144.47, @@ -2437,140 +2437,140 @@ "48": 145.25, "49": 144.24 }, - "expected_e2e_ms": 9298.75, + "expected_e2e_ms": 8612.26, "expected_avg_denoise_ms": 144.68, "expected_median_denoise_ms": 144.74, "estimated_full_test_time_s": 129.3 }, "wan2_1_t2v_1.3b_frame_interp_2x_upscaling_4x": { "stages_ms": { - "InputValidationStage": 0.07, - "TextEncodingStage": 1123.44, - "LatentPreparationStage": 0.16, + "InputValidationStage": 0.04, + "TextEncodingStage": 559.97, + "LatentPreparationStage": 0.14, "TimestepPreparationStage": 2.34, - "DenoisingStage": 7753.24, - "DecodingStage": 630.74 + "DenoisingStage": 7419.14, + "DecodingStage": 564.64 }, "denoise_step_ms": { - "0": 110.19, - "1": 154.98, - "2": 152.52, - "3": 154.48, - "4": 154.25, - "5": 156.71, + "0": 110.01, + "1": 125.73, + "2": 147.19, + "3": 145.54, + "4": 145.31, + "5": 146.31, "6": 158.16, - "7": 153.98, - "8": 154.7, - "9": 154.82, - "10": 154.0, - "11": 157.22, - "12": 156.51, + "7": 148.97, + "8": 146.24, + "9": 148.82, + "10": 145.96, + "11": 146.19, + "12": 151.66, "13": 154.95, - "14": 155.39, - "15": 156.07, - "16": 154.87, - "17": 157.5, - "18": 155.4, - "19": 157.49, - "20": 154.28, - "21": 155.94, - "22": 154.35, - "23": 156.62, - "24": 156.55, - "25": 154.99, - "26": 155.67, - "27": 156.44, - "28": 155.89, - "29": 156.77, - "30": 156.16, - "31": 154.89, - "32": 156.86, - "33": 155.82, - "34": 156.08, - "35": 157.5, - "36": 155.53, - "37": 155.38, - "38": 157.15, - "39": 156.91, - "40": 155.29, - "41": 157.02, - "42": 156.41, - "43": 157.29, - "44": 155.01, - "45": 156.07, - "46": 158.24, - "47": 155.49, - "48": 157.37, - "49": 155.54 + "14": 150.19, + "15": 146.77, + "16": 149.06, + "17": 146.95, + "18": 147.98, + "19": 153.21, + "20": 151.47, + "21": 147.68, + "22": 146.38, + "23": 148.52, + "24": 147.07, + "25": 151.59, + "26": 150.39, + "27": 150.41, + "28": 151.1, + "29": 148.02, + "30": 147.85, + "31": 148.52, + "32": 152.62, + "33": 150.97, + "34": 149.8, + "35": 150.53, + "36": 148.13, + "37": 148.54, + "38": 148.74, + "39": 149.68, + "40": 151.1, + "41": 150.29, + "42": 146.85, + "43": 146.56, + "44": 149.6, + "45": 148.65, + "46": 150.93, + "47": 150.19, + "48": 151.48, + "49": 147.91 }, - "expected_e2e_ms": 9918.3, - "expected_avg_denoise_ms": 154.95, - "expected_median_denoise_ms": 149.9, + "expected_e2e_ms": 8630.42, + "expected_avg_denoise_ms": 148.25, + "expected_median_denoise_ms": 148.78, "estimated_full_test_time_s": 129.4 }, "ltx_2.3_one_stage_ti2v": { "stages_ms": { "InputValidationStage": 3.6, - "TextEncodingStage": 1557.57, - "LTX2TextConnectorStage": 27.92, - "LTX2SigmaPreparationStage": 0.26, - "TimestepPreparationStage": 13.22, - "LTX2AVLatentPreparationStage": 0.17, + "TextEncodingStage": 773.87, + "LTX2TextConnectorStage": 27.35, + "LTX2SigmaPreparationStage": 0.19, + "TimestepPreparationStage": 11.85, + "LTX2AVLatentPreparationStage": 0.11, "LTX2ImageEncodingStage": 29.61, - "LTX2AVDenoisingStage": 15447.81, + "LTX2AVDenoisingStage": 15025.95, "LTX2AVDecodingStage": 234.96 }, "denoise_step_ms": { "0": 513.68, - "1": 497.17, - "2": 506.46, - "3": 510.93, - "4": 521.26, - "5": 601.18, - "6": 559.36, - "7": 514.61, - "8": 511.22, - "9": 507.85, - "10": 512.55, - "11": 500.76, - "12": 508.08, - "13": 574.14, - "14": 499.7, - "15": 499.14, - "16": 511.01, - "17": 503.34, - "18": 510.08, - "19": 527.73, - "20": 513.17, - "21": 504.58, - "22": 501.33, - "23": 498.04, - "24": 500.19, - "25": 505.21, - "26": 516.86, + "1": 493.96, + "2": 493.79, + "3": 495.7, + "4": 494.61, + "5": 495.44, + "6": 494.66, + "7": 493.61, + "8": 495.14, + "9": 494.39, + "10": 495.59, + "11": 493.94, + "12": 493.73, + "13": 495.33, + "14": 495.09, + "15": 493.96, + "16": 494.79, + "17": 494.07, + "18": 494.39, + "19": 495.85, + "20": 495.73, + "21": 495.74, + "22": 494.92, + "23": 494.89, + "24": 494.84, + "25": 494.45, + "26": 495.61, "27": 505.17, "28": 508.44, - "29": 499.97 + "29": 497.15 }, - "expected_e2e_ms": 17881.82, - "expected_avg_denoise_ms": 514.77, - "expected_median_denoise_ms": 508.26, + "expected_e2e_ms": 16722.42, + "expected_avg_denoise_ms": 500.63, + "expected_median_denoise_ms": 494.9, "estimated_full_test_time_s": 144.2 }, "ltx_2.3_two_stage_t2v_2gpus": { "stages_ms": { - "InputValidationStage": 0.07, - "TextEncodingStage": 1784.13, + "InputValidationStage": 0.05, + "TextEncodingStage": 775.72, "LTX2TextConnectorStage": 29.83, "LTX2HalveResolutionStage": 0.06, - "LTX2LoRASwitchStage": 153.74, - "LTX2SigmaPreparationStage": 0.63, - "TimestepPreparationStage": 31.17, - "LTX2AVLatentPreparationStage": 0.26, + "LTX2LoRASwitchStage": 0.01, + "LTX2SigmaPreparationStage": 0.61, + "TimestepPreparationStage": 20.77, + "LTX2AVLatentPreparationStage": 0.12, "LTX2ImageEncodingStage": 0.02, "LTX2AVDenoisingStage": 5460.05, "LTX2UpsampleStage": 3.31, - "LTX2RefinementStage": 615.01, + "LTX2RefinementStage": 589.32, "LTX2AVDecodingStage": 240.26, "per_frame_generation": null }, @@ -2588,87 +2588,87 @@ "10": 163.28, "11": 170.97, "12": 176.57, - "13": 189.92, - "14": 162.76, - "15": 179.89, + "13": 136.28, + "14": 138.62, + "15": 172.68, "16": 172.41, "17": 184.39, "18": 161.55, "19": 176.63, "20": 200.48, - "21": 188.38, + "21": 182.02, "22": 161.98, - "23": 172.25, - "24": 167.45, - "25": 189.81, - "26": 176.08, - "27": 170.88, - "28": 188.96, + "23": 139.58, + "24": 137.21, + "25": 182.74, + "26": 149.38, + "27": 144.21, + "28": 171.22, "29": 183.21, - "30": 238.92, - "31": 187.04, - "32": 184.4 + "30": 142.41, + "31": 134.02, + "32": 138.32 }, - "expected_e2e_ms": 18039.38, - "expected_avg_denoise_ms": 271.95, + "expected_e2e_ms": 12925.1, + "expected_avg_denoise_ms": 209.57, "expected_median_denoise_ms": 179.26, "estimated_full_test_time_s": 160.0 }, "ltx_2_3_two_stage_ti2v_2gpus": { "stages_ms": { "InputValidationStage": 3.21, - "TextEncodingStage": 1554.35, - "LTX2TextConnectorStage": 27.64, + "TextEncodingStage": 747.28, + "LTX2TextConnectorStage": 27.37, "LTX2HalveResolutionStage": 0.05, - "LTX2LoRASwitchStage": 70.0, - "LTX2SigmaPreparationStage": 0.4, - "TimestepPreparationStage": 23.51, - "LTX2AVLatentPreparationStage": 0.15, + "LTX2LoRASwitchStage": 3.01, + "LTX2SigmaPreparationStage": 0.22, + "TimestepPreparationStage": 19.94, + "LTX2AVLatentPreparationStage": 0.14, "LTX2ImageEncodingStage": 27.58, "LTX2AVDenoisingStage": 7506.38, "LTX2UpsampleStage": 2.31, - "LTX2RefinementStage": 688.98, + "LTX2RefinementStage": 670.53, "LTX2AVDecodingStage": 250.93, "per_frame_generation": null }, "denoise_step_ms": { - "0": 280.97, - "1": 274.21, + "0": 261.94, + "1": 271.84, "2": 250.26, "3": 251.97, - "4": 251.58, - "5": 253.73, - "6": 250.99, - "7": 244.5, + "4": 245.27, + "5": 244.88, + "6": 225.68, + "7": 243.6, "8": 249.33, "9": 246.74, "10": 245.15, - "11": 253.89, - "12": 260.57, + "11": 246.62, + "12": 245.31, "13": 238.39, "14": 244.41, "15": 242.77, "16": 246.85, "17": 245.53, "18": 246.1, - "19": 247.94, - "20": 251.52, - "21": 271.21, - "22": 249.64, + "19": 246.07, + "20": 247.84, + "21": 250.74, + "22": 249.19, "23": 246.86, "24": 239.52, "25": 249.31, - "26": 245.69, - "27": 246.51, + "26": 242.79, + "27": 246.48, "28": 238.94, "29": 237.2, "30": 127.08, "31": 123.1, - "32": 209.80 + "32": 193.15 }, - "expected_e2e_ms": 12000.0, - "expected_avg_denoise_ms": 299.01, - "expected_median_denoise_ms": 246.85, + "expected_e2e_ms": 9687.04, + "expected_avg_denoise_ms": 250.16, + "expected_median_denoise_ms": 242.79, "estimated_full_test_time_s": 170.0 }, "lingbot_world_realtime_plastic_beach": { @@ -2681,44 +2681,44 @@ }, "ltx_2_3_hq_pipeline": { "stages_ms": { - "InputValidationStage": 0.08, - "TextEncodingStage": 803.8, - "LTX2TextConnectorStage": 31.22, - "LTX2HalveResolutionStage": 0.06, - "LTX2LoRASwitchStage": 0.01, - "LTX2SigmaPreparationStage": 0.27, - "TimestepPreparationStage": 60.49, - "LTX2AVLatentPreparationStage": 0.18, - "LTX2ImageEncodingStage": 1185.13, - "LTX2AVDenoisingStage": 22802.53, - "LTX2UpsampleStage": 37.42, - "LTX2RefinementStage": 3123.85, - "LTX2AVDecodingStage": 3266.55, + "InputValidationStage": 11.84, + "TextEncodingStage": 446.77, + "LTX2TextConnectorStage": 32.96, + "LTX2HalveResolutionStage": 0.07, + "LTX2LoRASwitchStage": 0.03, + "LTX2SigmaPreparationStage": 0.38, + "TimestepPreparationStage": 24.67, + "LTX2AVLatentPreparationStage": 0.16, + "LTX2ImageEncodingStage": 1133.15, + "LTX2AVDenoisingStage": 19539.95, + "LTX2UpsampleStage": 22.73, + "LTX2RefinementStage": 3059.95, + "LTX2AVDecodingStage": 1744.05, "per_frame_generation": null }, "denoise_step_ms": { - "0": 1942.27, - "1": 1736.24, - "2": 1740.39, - "3": 1816.87, - "4": 1576.27, - "5": 1611.1, - "6": 1586.7, - "7": 1559.69, - "8": 1456.36, - "9": 1591.21, - "10": 1534.93, - "11": 852.43, - "12": 836.03, - "13": 855.98, - "14": 443.03, - "15": 957.19, + "0": 1677.39, + "1": 778.3, + "2": 740.93, + "3": 1177.58, + "4": 751.77, + "5": 744.01, + "6": 742.02, + "7": 1264.88, + "8": 744.11, + "9": 739.84, + "10": 1209.91, + "11": 739.62, + "12": 736.7, + "13": 733.55, + "14": 590.44, + "15": 903.04, "16": 876.98, - "17": 878.4 + "17": 879.42 }, - "expected_e2e_ms": 30068.97, - "expected_avg_denoise_ms": 1325.11, - "expected_median_denoise_ms": 1547.31, + "expected_e2e_ms": 27309.99, + "expected_avg_denoise_ms": 1159.69, + "expected_median_denoise_ms": 1199, "estimated_full_test_time_s": 363.2 }, "qwen_image_t2i_cache_dit_scm_config_diffusers_1gpu": { @@ -2784,9 +2784,9 @@ "stages_ms": { "Cosmos3ImagePreprocessStage": 0.01, "Cosmos3TokenizationStage": 152.97, - "Cosmos3LatentPreparationStage": 0.17, - "Cosmos3TimestepPreparationStage": 0.64, - "Cosmos3DenoisingStage": 294.92, + "Cosmos3LatentPreparationStage": 0.14, + "Cosmos3TimestepPreparationStage": 0.56, + "Cosmos3DenoisingStage": 261.27, "Cosmos3DecodingStage": 47.82, "per_frame_generation": null }, @@ -2796,7 +2796,7 @@ "2": 85.06, "3": 58.08 }, - "expected_e2e_ms": 499.70, + "expected_e2e_ms": 499.7, "expected_avg_denoise_ms": 73.47, "expected_median_denoise_ms": 73.29, "estimated_full_test_time_s": 65.0