Datasets:

hexuan21

/

VideoFeedback_r2k_bad8k

like 0

[ "7001354_00.jpg", "7001354_01.jpg", "7001354_02.jpg", "7001354_03.jpg", "7001354_04.jpg", "7001354_05.jpg", "7001354_06.jpg", "7001354_07.jpg" ]

[ { "from": "human", "value": "\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos,\nplease watch the following frames of a given video and see the text prompt for generating the video,\nthen give scores from 5 different dimensions:\n(1) visual quality: the quality of the video in terms of clearness, resolution, brightness, and color\n(2) temporal consistency, both the consistency of objects or humans and the smoothness of motion or movements\n(3) dynamic degree, the degree of dynamic changes\n(4) text-to-video alignment, the alignment between the text prompt and the video content\n(5) factual consistency, the consistency of the video content with the common-sense and factual knowledge\n\nfor each dimension, output a float number from 1.0 to 4.0,\nthe higher the number is, the better the video performs in that sub-score, \nthe lowest 1.0 means Bad, the highest 4.0 means Perfect/Real (the video is like a real video)\nHere is an output example:\nvisual quality: 3.2\ntemporal consistency: 2.7\ndynamic degree: 4.0\ntext-to-video alignment: 2.3\nfactual consistency: 1.8\n\nFor this video, the text prompt is \"A scenary of a tv show about enginery\",\nall the frames of video are as follows:\n\n<image> <image> <image> <image> <image> <image> <image> <image> " }, { "from": "gpt", "value": "visual quality: 2\n temporal consistency: 2\n dynamic degree: 3\n text-to-video alignment: 3\n factual consistency: 3\n" } ]

[ "6001550_00.jpg", "6001550_01.jpg", "6001550_02.jpg", "6001550_03.jpg", "6001550_04.jpg", "6001550_05.jpg", "6001550_06.jpg", "6001550_07.jpg", "6001550_08.jpg", "6001550_09.jpg", "6001550_10.jpg", "6001550_11.jpg", "6001550_12.jpg", "6001550_13.jpg", "6001550_14.jpg", "6001550_15.jpg" ]

[ { "from": "human", "value": "\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos,\nplease watch the following frames of a given video and see the text prompt for generating the video,\nthen give scores from 5 different dimensions:\n(1) visual quality: the quality of the video in terms of clearness, resolution, brightness, and color\n(2) temporal consistency, both the consistency of objects or humans and the smoothness of motion or movements\n(3) dynamic degree, the degree of dynamic changes\n(4) text-to-video alignment, the alignment between the text prompt and the video content\n(5) factual consistency, the consistency of the video content with the common-sense and factual knowledge\n\nfor each dimension, output a float number from 1.0 to 4.0,\nthe higher the number is, the better the video performs in that sub-score, \nthe lowest 1.0 means Bad, the highest 4.0 means Perfect/Real (the video is like a real video)\nHere is an output example:\nvisual quality: 3.2\ntemporal consistency: 2.7\ndynamic degree: 4.0\ntext-to-video alignment: 2.3\nfactual consistency: 1.8\n\nFor this video, the text prompt is \"Conservationists working to protect cheetahs in their natural habitat\",\nall the frames of video are as follows:\n\n<image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> " }, { "from": "gpt", "value": "visual quality: 1\n temporal consistency: 2\n dynamic degree: 3\n text-to-video alignment: 3\n factual consistency: 1\n" } ]

[ "1000514_00.jpg", "1000514_01.jpg", "1000514_02.jpg", "1000514_03.jpg", "1000514_04.jpg", "1000514_05.jpg", "1000514_06.jpg", "1000514_07.jpg", "1000514_08.jpg", "1000514_09.jpg", "1000514_10.jpg", "1000514_11.jpg", "1000514_12.jpg", "1000514_13.jpg", "1000514_14.jpg", "1000514_15.jpg" ]

[ { "from": "human", "value": "\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos,\nplease watch the following frames of a given video and see the text prompt for generating the video,\nthen give scores from 5 different dimensions:\n(1) visual quality: the quality of the video in terms of clearness, resolution, brightness, and color\n(2) temporal consistency, both the consistency of objects or humans and the smoothness of motion or movements\n(3) dynamic degree, the degree of dynamic changes\n(4) text-to-video alignment, the alignment between the text prompt and the video content\n(5) factual consistency, the consistency of the video content with the common-sense and factual knowledge\n\nfor each dimension, output a float number from 1.0 to 4.0,\nthe higher the number is, the better the video performs in that sub-score, \nthe lowest 1.0 means Bad, the highest 4.0 means Perfect/Real (the video is like a real video)\nHere is an output example:\nvisual quality: 3.2\ntemporal consistency: 2.7\ndynamic degree: 4.0\ntext-to-video alignment: 2.3\nfactual consistency: 1.8\n\nFor this video, the text prompt is \"The rocket destroy the Moscow Kremlin\",\nall the frames of video are as follows:\n\n<image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> " }, { "from": "gpt", "value": "visual quality: 1\n temporal consistency: 2\n dynamic degree: 3\n text-to-video alignment: 2\n factual consistency: 1\n" } ]

[ "1002316_00.jpg", "1002316_01.jpg", "1002316_02.jpg", "1002316_03.jpg", "1002316_04.jpg", "1002316_05.jpg", "1002316_06.jpg", "1002316_07.jpg", "1002316_08.jpg", "1002316_09.jpg", "1002316_10.jpg", "1002316_11.jpg", "1002316_12.jpg", "1002316_13.jpg", "1002316_14.jpg", "1002316_15.jpg" ]

[ { "from": "human", "value": "\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos,\nplease watch the following frames of a given video and see the text prompt for generating the video,\nthen give scores from 5 different dimensions:\n(1) visual quality: the quality of the video in terms of clearness, resolution, brightness, and color\n(2) temporal consistency, both the consistency of objects or humans and the smoothness of motion or movements\n(3) dynamic degree, the degree of dynamic changes\n(4) text-to-video alignment, the alignment between the text prompt and the video content\n(5) factual consistency, the consistency of the video content with the common-sense and factual knowledge\n\nfor each dimension, output a float number from 1.0 to 4.0,\nthe higher the number is, the better the video performs in that sub-score, \nthe lowest 1.0 means Bad, the highest 4.0 means Perfect/Real (the video is like a real video)\nHere is an output example:\nvisual quality: 3.2\ntemporal consistency: 2.7\ndynamic degree: 4.0\ntext-to-video alignment: 2.3\nfactual consistency: 1.8\n\nFor this video, the text prompt is \"She Quotes Looking at Me Silently, Smiling Mesmerizingly, Film, High Detail, Well Exposed, Theme Sound, Dynamic Movement\",\nall the frames of video are as follows:\n\n<image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> " }, { "from": "gpt", "value": "visual quality: 1\n temporal consistency: 2\n dynamic degree: 3\n text-to-video alignment: 2\n factual consistency: 1\n" } ]

[ "3000942_00.jpg", "3000942_01.jpg", "3000942_02.jpg", "3000942_03.jpg", "3000942_04.jpg", "3000942_05.jpg", "3000942_06.jpg", "3000942_07.jpg", "3000942_08.jpg", "3000942_09.jpg", "3000942_10.jpg", "3000942_11.jpg", "3000942_12.jpg", "3000942_13.jpg", "3000942_14.jpg", "3000942_15.jpg" ]

[ { "from": "human", "value": "\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos,\nplease watch the following frames of a given video and see the text prompt for generating the video,\nthen give scores from 5 different dimensions:\n(1) visual quality: the quality of the video in terms of clearness, resolution, brightness, and color\n(2) temporal consistency, both the consistency of objects or humans and the smoothness of motion or movements\n(3) dynamic degree, the degree of dynamic changes\n(4) text-to-video alignment, the alignment between the text prompt and the video content\n(5) factual consistency, the consistency of the video content with the common-sense and factual knowledge\n\nfor each dimension, output a float number from 1.0 to 4.0,\nthe higher the number is, the better the video performs in that sub-score, \nthe lowest 1.0 means Bad, the highest 4.0 means Perfect/Real (the video is like a real video)\nHere is an output example:\nvisual quality: 3.2\ntemporal consistency: 2.7\ndynamic degree: 4.0\ntext-to-video alignment: 2.3\nfactual consistency: 1.8\n\nFor this video, the text prompt is \"A stoic statue of a philosopher in a majestic pose, with crossed arms and a determined and firm gaze\",\nall the frames of video are as follows:\n\n<image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> " }, { "from": "gpt", "value": "visual quality: 2\n temporal consistency: 3\n dynamic degree: 1\n text-to-video alignment: 3\n factual consistency: 3\n" } ]

[ "3000791_00.jpg", "3000791_01.jpg", "3000791_02.jpg", "3000791_03.jpg", "3000791_04.jpg", "3000791_05.jpg", "3000791_06.jpg", "3000791_07.jpg", "3000791_08.jpg", "3000791_09.jpg", "3000791_10.jpg", "3000791_11.jpg", "3000791_12.jpg", "3000791_13.jpg", "3000791_14.jpg", "3000791_15.jpg" ]

[ { "from": "human", "value": "\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos,\nplease watch the following frames of a given video and see the text prompt for generating the video,\nthen give scores from 5 different dimensions:\n(1) visual quality: the quality of the video in terms of clearness, resolution, brightness, and color\n(2) temporal consistency, both the consistency of objects or humans and the smoothness of motion or movements\n(3) dynamic degree, the degree of dynamic changes\n(4) text-to-video alignment, the alignment between the text prompt and the video content\n(5) factual consistency, the consistency of the video content with the common-sense and factual knowledge\n\nfor each dimension, output a float number from 1.0 to 4.0,\nthe higher the number is, the better the video performs in that sub-score, \nthe lowest 1.0 means Bad, the highest 4.0 means Perfect/Real (the video is like a real video)\nHere is an output example:\nvisual quality: 3.2\ntemporal consistency: 2.7\ndynamic degree: 4.0\ntext-to-video alignment: 2.3\nfactual consistency: 1.8\n\nFor this video, the text prompt is \"make bjork president of the united states\",\nall the frames of video are as follows:\n\n<image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> " }, { "from": "gpt", "value": "visual quality: 3\n temporal consistency: 2\n dynamic degree: 3\n text-to-video alignment: 2\n factual consistency: 2\n" } ]

[ "1001518_00.jpg", "1001518_01.jpg", "1001518_02.jpg", "1001518_03.jpg", "1001518_04.jpg", "1001518_05.jpg", "1001518_06.jpg", "1001518_07.jpg", "1001518_08.jpg", "1001518_09.jpg", "1001518_10.jpg", "1001518_11.jpg", "1001518_12.jpg", "1001518_13.jpg", "1001518_14.jpg", "1001518_15.jpg" ]

[ { "from": "human", "value": "\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos,\nplease watch the following frames of a given video and see the text prompt for generating the video,\nthen give scores from 5 different dimensions:\n(1) visual quality: the quality of the video in terms of clearness, resolution, brightness, and color\n(2) temporal consistency, both the consistency of objects or humans and the smoothness of motion or movements\n(3) dynamic degree, the degree of dynamic changes\n(4) text-to-video alignment, the alignment between the text prompt and the video content\n(5) factual consistency, the consistency of the video content with the common-sense and factual knowledge\n\nfor each dimension, output a float number from 1.0 to 4.0,\nthe higher the number is, the better the video performs in that sub-score, \nthe lowest 1.0 means Bad, the highest 4.0 means Perfect/Real (the video is like a real video)\nHere is an output example:\nvisual quality: 3.2\ntemporal consistency: 2.7\ndynamic degree: 4.0\ntext-to-video alignment: 2.3\nfactual consistency: 1.8\n\nFor this video, the text prompt is \"cinematic video footage of a white rabbit with a black tie hiding within a tree trunk, magical realism\",\nall the frames of video are as follows:\n\n<image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> " }, { "from": "gpt", "value": "visual quality: 1\n temporal consistency: 1\n dynamic degree: 3\n text-to-video alignment: 2\n factual consistency: 1\n" } ]

[ "4001591_00.jpg", "4001591_01.jpg", "4001591_02.jpg", "4001591_03.jpg", "4001591_04.jpg", "4001591_05.jpg", "4001591_06.jpg", "4001591_07.jpg", "4001591_08.jpg", "4001591_09.jpg", "4001591_10.jpg", "4001591_11.jpg", "4001591_12.jpg", "4001591_13.jpg", "4001591_14.jpg", "4001591_15.jpg" ]

[ { "from": "human", "value": "\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos,\nplease watch the following frames of a given video and see the text prompt for generating the video,\nthen give scores from 5 different dimensions:\n(1) visual quality: the quality of the video in terms of clearness, resolution, brightness, and color\n(2) temporal consistency, both the consistency of objects or humans and the smoothness of motion or movements\n(3) dynamic degree, the degree of dynamic changes\n(4) text-to-video alignment, the alignment between the text prompt and the video content\n(5) factual consistency, the consistency of the video content with the common-sense and factual knowledge\n\nfor each dimension, output a float number from 1.0 to 4.0,\nthe higher the number is, the better the video performs in that sub-score, \nthe lowest 1.0 means Bad, the highest 4.0 means Perfect/Real (the video is like a real video)\nHere is an output example:\nvisual quality: 3.2\ntemporal consistency: 2.7\ndynamic degree: 4.0\ntext-to-video alignment: 2.3\nfactual consistency: 1.8\n\nFor this video, the text prompt is \"indian god Shiva playing video games\",\nall the frames of video are as follows:\n\n<image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> " }, { "from": "gpt", "value": "visual quality: 3\n temporal consistency: 2\n dynamic degree: 1\n text-to-video alignment: 3\n factual consistency: 3\n" } ]

[ "8000180_00.jpg", "8000180_01.jpg", "8000180_02.jpg", "8000180_03.jpg", "8000180_04.jpg", "8000180_05.jpg", "8000180_06.jpg", "8000180_07.jpg", "8000180_08.jpg", "8000180_09.jpg", "8000180_10.jpg", "8000180_11.jpg", "8000180_12.jpg", "8000180_13.jpg", "8000180_14.jpg", "8000180_15.jpg", "8000180_16.jpg", "8000180_17.jpg", "8000180_18.jpg", "8000180_19.jpg", "8000180_20.jpg", "8000180_21.jpg", "8000180_22.jpg", "8000180_23.jpg" ]

[ { "from": "human", "value": "\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos,\nplease watch the following frames of a given video and see the text prompt for generating the video,\nthen give scores from 5 different dimensions:\n(1) visual quality: the quality of the video in terms of clearness, resolution, brightness, and color\n(2) temporal consistency, both the consistency of objects or humans and the smoothness of motion or movements\n(3) dynamic degree, the degree of dynamic changes\n(4) text-to-video alignment, the alignment between the text prompt and the video content\n(5) factual consistency, the consistency of the video content with the common-sense and factual knowledge\n\nfor each dimension, output a float number from 1.0 to 4.0,\nthe higher the number is, the better the video performs in that sub-score, \nthe lowest 1.0 means Bad, the highest 4.0 means Perfect/Real (the video is like a real video)\nHere is an output example:\nvisual quality: 3.2\ntemporal consistency: 2.7\ndynamic degree: 4.0\ntext-to-video alignment: 2.3\nfactual consistency: 1.8\n\nFor this video, the text prompt is \"a cyber punk AI machine in the style of tim burton putting on a pair of glasses in front of a face recognition camera and falsely getting accepted for another person due to fooling the system for being a human while devilishly smiling\",\nall the frames of video are as follows:\n\n<image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> " }, { "from": "gpt", "value": "visual quality: 1\n temporal consistency: 1\n dynamic degree: 3\n text-to-video alignment: 2\n factual consistency: 1\n" } ]

[ "3001012_00.jpg", "3001012_01.jpg", "3001012_02.jpg", "3001012_03.jpg", "3001012_04.jpg", "3001012_05.jpg", "3001012_06.jpg", "3001012_07.jpg", "3001012_08.jpg", "3001012_09.jpg", "3001012_10.jpg", "3001012_11.jpg", "3001012_12.jpg", "3001012_13.jpg", "3001012_14.jpg", "3001012_15.jpg" ]

[ { "from": "human", "value": "\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos,\nplease watch the following frames of a given video and see the text prompt for generating the video,\nthen give scores from 5 different dimensions:\n(1) visual quality: the quality of the video in terms of clearness, resolution, brightness, and color\n(2) temporal consistency, both the consistency of objects or humans and the smoothness of motion or movements\n(3) dynamic degree, the degree of dynamic changes\n(4) text-to-video alignment, the alignment between the text prompt and the video content\n(5) factual consistency, the consistency of the video content with the common-sense and factual knowledge\n\nfor each dimension, output a float number from 1.0 to 4.0,\nthe higher the number is, the better the video performs in that sub-score, \nthe lowest 1.0 means Bad, the highest 4.0 means Perfect/Real (the video is like a real video)\nHere is an output example:\nvisual quality: 3.2\ntemporal consistency: 2.7\ndynamic degree: 4.0\ntext-to-video alignment: 2.3\nfactual consistency: 1.8\n\nFor this video, the text prompt is \"Once upon a time in a quaint village nestled between rolling hills, there lived a wise old gardener named Ezra.\",\nall the frames of video are as follows:\n\n<image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> " }, { "from": "gpt", "value": "visual quality: 2\n temporal consistency: 1\n dynamic degree: 3\n text-to-video alignment: 2\n factual consistency: 2\n" } ]

[ "5000038_00.jpg", "5000038_01.jpg", "5000038_02.jpg", "5000038_03.jpg", "5000038_04.jpg", "5000038_05.jpg", "5000038_06.jpg", "5000038_07.jpg", "5000038_08.jpg", "5000038_09.jpg", "5000038_10.jpg", "5000038_11.jpg", "5000038_12.jpg", "5000038_13.jpg", "5000038_14.jpg", "5000038_15.jpg" ]

[ { "from": "human", "value": "\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos,\nplease watch the following frames of a given video and see the text prompt for generating the video,\nthen give scores from 5 different dimensions:\n(1) visual quality: the quality of the video in terms of clearness, resolution, brightness, and color\n(2) temporal consistency, both the consistency of objects or humans and the smoothness of motion or movements\n(3) dynamic degree, the degree of dynamic changes\n(4) text-to-video alignment, the alignment between the text prompt and the video content\n(5) factual consistency, the consistency of the video content with the common-sense and factual knowledge\n\nfor each dimension, output a float number from 1.0 to 4.0,\nthe higher the number is, the better the video performs in that sub-score, \nthe lowest 1.0 means Bad, the highest 4.0 means Perfect/Real (the video is like a real video)\nHere is an output example:\nvisual quality: 3.2\ntemporal consistency: 2.7\ndynamic degree: 4.0\ntext-to-video alignment: 2.3\nfactual consistency: 1.8\n\nFor this video, the text prompt is \"G Man skibidi toilet attack the Titan Tv man.\",\nall the frames of video are as follows:\n\n<image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> " }, { "from": "gpt", "value": "visual quality: 2\n temporal consistency: 1\n dynamic degree: 3\n text-to-video alignment: 3\n factual consistency: 1\n" } ]

[ "5000282_00.jpg", "5000282_01.jpg", "5000282_02.jpg", "5000282_03.jpg", "5000282_04.jpg", "5000282_05.jpg", "5000282_06.jpg", "5000282_07.jpg", "5000282_08.jpg", "5000282_09.jpg", "5000282_10.jpg", "5000282_11.jpg", "5000282_12.jpg", "5000282_13.jpg", "5000282_14.jpg", "5000282_15.jpg" ]

[ { "from": "human", "value": "\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos,\nplease watch the following frames of a given video and see the text prompt for generating the video,\nthen give scores from 5 different dimensions:\n(1) visual quality: the quality of the video in terms of clearness, resolution, brightness, and color\n(2) temporal consistency, both the consistency of objects or humans and the smoothness of motion or movements\n(3) dynamic degree, the degree of dynamic changes\n(4) text-to-video alignment, the alignment between the text prompt and the video content\n(5) factual consistency, the consistency of the video content with the common-sense and factual knowledge\n\nfor each dimension, output a float number from 1.0 to 4.0,\nthe higher the number is, the better the video performs in that sub-score, \nthe lowest 1.0 means Bad, the highest 4.0 means Perfect/Real (the video is like a real video)\nHere is an output example:\nvisual quality: 3.2\ntemporal consistency: 2.7\ndynamic degree: 4.0\ntext-to-video alignment: 2.3\nfactual consistency: 1.8\n\nFor this video, the text prompt is \"Chubby woman walking. Gym uniform, back view, wearing navy blue bloomers,\",\nall the frames of video are as follows:\n\n<image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> " }, { "from": "gpt", "value": "visual quality: 3\n temporal consistency: 1\n dynamic degree: 3\n text-to-video alignment: 3\n factual consistency: 1\n" } ]

[ "0000372_00.jpg", "0000372_01.jpg", "0000372_02.jpg", "0000372_03.jpg", "0000372_04.jpg", "0000372_05.jpg", "0000372_06.jpg", "0000372_07.jpg", "0000372_08.jpg", "0000372_09.jpg", "0000372_10.jpg", "0000372_11.jpg", "0000372_12.jpg", "0000372_13.jpg", "0000372_14.jpg", "0000372_15.jpg", "0000372_16.jpg", "0000372_17.jpg", "0000372_18.jpg", "0000372_19.jpg", "0000372_20.jpg", "0000372_21.jpg", "0000372_22.jpg", "0000372_23.jpg" ]

[ { "from": "human", "value": "\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos,\nplease watch the following frames of a given video and see the text prompt for generating the video,\nthen give scores from 5 different dimensions:\n(1) visual quality: the quality of the video in terms of clearness, resolution, brightness, and color\n(2) temporal consistency, both the consistency of objects or humans and the smoothness of motion or movements\n(3) dynamic degree, the degree of dynamic changes\n(4) text-to-video alignment, the alignment between the text prompt and the video content\n(5) factual consistency, the consistency of the video content with the common-sense and factual knowledge\n\nfor each dimension, output a float number from 1.0 to 4.0,\nthe higher the number is, the better the video performs in that sub-score, \nthe lowest 1.0 means Bad, the highest 4.0 means Perfect/Real (the video is like a real video)\nHere is an output example:\nvisual quality: 3.2\ntemporal consistency: 2.7\ndynamic degree: 4.0\ntext-to-video alignment: 2.3\nfactual consistency: 1.8\n\nFor this video, the text prompt is \"tom cruise smiling and blinking one eye\",\nall the frames of video are as follows:\n\n<image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> " }, { "from": "gpt", "value": "visual quality: 4\n temporal consistency: 4\n dynamic degree: 4\n text-to-video alignment: 4\n factual consistency: 4\n" } ]

[ "3002115_00.jpg", "3002115_01.jpg", "3002115_02.jpg", "3002115_03.jpg", "3002115_04.jpg", "3002115_05.jpg", "3002115_06.jpg", "3002115_07.jpg", "3002115_08.jpg", "3002115_09.jpg", "3002115_10.jpg", "3002115_11.jpg", "3002115_12.jpg", "3002115_13.jpg", "3002115_14.jpg", "3002115_15.jpg" ]

[ { "from": "human", "value": "\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos,\nplease watch the following frames of a given video and see the text prompt for generating the video,\nthen give scores from 5 different dimensions:\n(1) visual quality: the quality of the video in terms of clearness, resolution, brightness, and color\n(2) temporal consistency, both the consistency of objects or humans and the smoothness of motion or movements\n(3) dynamic degree, the degree of dynamic changes\n(4) text-to-video alignment, the alignment between the text prompt and the video content\n(5) factual consistency, the consistency of the video content with the common-sense and factual knowledge\n\nfor each dimension, output a float number from 1.0 to 4.0,\nthe higher the number is, the better the video performs in that sub-score, \nthe lowest 1.0 means Bad, the highest 4.0 means Perfect/Real (the video is like a real video)\nHere is an output example:\nvisual quality: 3.2\ntemporal consistency: 2.7\ndynamic degree: 4.0\ntext-to-video alignment: 2.3\nfactual consistency: 1.8\n\nFor this video, the text prompt is \"anime man holding a mobile phone in call , back camera , futuristic aesthetics\",\nall the frames of video are as follows:\n\n<image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> " }, { "from": "gpt", "value": "visual quality: 2\n temporal consistency: 3\n dynamic degree: 1\n text-to-video alignment: 3\n factual consistency: 2\n" } ]

[ "6001645_00.jpg", "6001645_01.jpg", "6001645_02.jpg", "6001645_03.jpg", "6001645_04.jpg", "6001645_05.jpg", "6001645_06.jpg", "6001645_07.jpg", "6001645_08.jpg", "6001645_09.jpg", "6001645_10.jpg", "6001645_11.jpg", "6001645_12.jpg", "6001645_13.jpg", "6001645_14.jpg", "6001645_15.jpg" ]

[ { "from": "human", "value": "\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos,\nplease watch the following frames of a given video and see the text prompt for generating the video,\nthen give scores from 5 different dimensions:\n(1) visual quality: the quality of the video in terms of clearness, resolution, brightness, and color\n(2) temporal consistency, both the consistency of objects or humans and the smoothness of motion or movements\n(3) dynamic degree, the degree of dynamic changes\n(4) text-to-video alignment, the alignment between the text prompt and the video content\n(5) factual consistency, the consistency of the video content with the common-sense and factual knowledge\n\nfor each dimension, output a float number from 1.0 to 4.0,\nthe higher the number is, the better the video performs in that sub-score, \nthe lowest 1.0 means Bad, the highest 4.0 means Perfect/Real (the video is like a real video)\nHere is an output example:\nvisual quality: 3.2\ntemporal consistency: 2.7\ndynamic degree: 4.0\ntext-to-video alignment: 2.3\nfactual consistency: 1.8\n\nFor this video, the text prompt is \"a gigantic crowd has gathered in front of the State Duma, the boundaries of this crowd are not visible\",\nall the frames of video are as follows:\n\n<image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> " }, { "from": "gpt", "value": "visual quality: 1\n temporal consistency: 2\n dynamic degree: 3\n text-to-video alignment: 3\n factual consistency: 1\n" } ]

[ "2001874_00.jpg", "2001874_01.jpg", "2001874_02.jpg", "2001874_03.jpg", "2001874_04.jpg", "2001874_05.jpg", "2001874_06.jpg", "2001874_07.jpg", "2001874_08.jpg", "2001874_09.jpg", "2001874_10.jpg", "2001874_11.jpg", "2001874_12.jpg", "2001874_13.jpg", "2001874_14.jpg", "2001874_15.jpg" ]

[ { "from": "human", "value": "\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos,\nplease watch the following frames of a given video and see the text prompt for generating the video,\nthen give scores from 5 different dimensions:\n(1) visual quality: the quality of the video in terms of clearness, resolution, brightness, and color\n(2) temporal consistency, both the consistency of objects or humans and the smoothness of motion or movements\n(3) dynamic degree, the degree of dynamic changes\n(4) text-to-video alignment, the alignment between the text prompt and the video content\n(5) factual consistency, the consistency of the video content with the common-sense and factual knowledge\n\nfor each dimension, output a float number from 1.0 to 4.0,\nthe higher the number is, the better the video performs in that sub-score, \nthe lowest 1.0 means Bad, the highest 4.0 means Perfect/Real (the video is like a real video)\nHere is an output example:\nvisual quality: 3.2\ntemporal consistency: 2.7\ndynamic degree: 4.0\ntext-to-video alignment: 2.3\nfactual consistency: 1.8\n\nFor this video, the text prompt is \"Museum of History. The skeleton of a fossilized Basilosaurus is suspended in the center of the museum. You can see his shiny teeth and matte white bones.\",\nall the frames of video are as follows:\n\n<image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> " }, { "from": "gpt", "value": "visual quality: 3\n temporal consistency: 3\n dynamic degree: 1\n text-to-video alignment: 2\n factual consistency: 3\n" } ]

[ "0002042_00.jpg", "0002042_01.jpg", "0002042_02.jpg", "0002042_03.jpg", "0002042_04.jpg", "0002042_05.jpg", "0002042_06.jpg", "0002042_07.jpg", "0002042_08.jpg", "0002042_09.jpg", "0002042_10.jpg", "0002042_11.jpg", "0002042_12.jpg", "0002042_13.jpg", "0002042_14.jpg", "0002042_15.jpg", "0002042_16.jpg", "0002042_17.jpg", "0002042_18.jpg", "0002042_19.jpg", "0002042_20.jpg", "0002042_21.jpg", "0002042_22.jpg", "0002042_23.jpg" ]

[ { "from": "human", "value": "\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos,\nplease watch the following frames of a given video and see the text prompt for generating the video,\nthen give scores from 5 different dimensions:\n(1) visual quality: the quality of the video in terms of clearness, resolution, brightness, and color\n(2) temporal consistency, both the consistency of objects or humans and the smoothness of motion or movements\n(3) dynamic degree, the degree of dynamic changes\n(4) text-to-video alignment, the alignment between the text prompt and the video content\n(5) factual consistency, the consistency of the video content with the common-sense and factual knowledge\n\nfor each dimension, output a float number from 1.0 to 4.0,\nthe higher the number is, the better the video performs in that sub-score, \nthe lowest 1.0 means Bad, the highest 4.0 means Perfect/Real (the video is like a real video)\nHere is an output example:\nvisual quality: 3.2\ntemporal consistency: 2.7\ndynamic degree: 4.0\ntext-to-video alignment: 2.3\nfactual consistency: 1.8\n\nFor this video, the text prompt is \"A holographic globe spins in the center, surrounded by floating, interactive screens displaying the Socrates app interface.\",\nall the frames of video are as follows:\n\n<image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> " }, { "from": "gpt", "value": "visual quality: 2\n temporal consistency: 2\n dynamic degree: 2\n text-to-video alignment: 2\n factual consistency: 3\n" } ]

[ "3001227_00.jpg", "3001227_01.jpg", "3001227_02.jpg", "3001227_03.jpg", "3001227_04.jpg", "3001227_05.jpg", "3001227_06.jpg", "3001227_07.jpg", "3001227_08.jpg", "3001227_09.jpg", "3001227_10.jpg", "3001227_11.jpg", "3001227_12.jpg", "3001227_13.jpg", "3001227_14.jpg", "3001227_15.jpg" ]

[ { "from": "human", "value": "\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos,\nplease watch the following frames of a given video and see the text prompt for generating the video,\nthen give scores from 5 different dimensions:\n(1) visual quality: the quality of the video in terms of clearness, resolution, brightness, and color\n(2) temporal consistency, both the consistency of objects or humans and the smoothness of motion or movements\n(3) dynamic degree, the degree of dynamic changes\n(4) text-to-video alignment, the alignment between the text prompt and the video content\n(5) factual consistency, the consistency of the video content with the common-sense and factual knowledge\n\nfor each dimension, output a float number from 1.0 to 4.0,\nthe higher the number is, the better the video performs in that sub-score, \nthe lowest 1.0 means Bad, the highest 4.0 means Perfect/Real (the video is like a real video)\nHere is an output example:\nvisual quality: 3.2\ntemporal consistency: 2.7\ndynamic degree: 4.0\ntext-to-video alignment: 2.3\nfactual consistency: 1.8\n\nFor this video, the text prompt is \"space ship, medium full side angle, ar 16:9, flying speed light\",\nall the frames of video are as follows:\n\n<image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> " }, { "from": "gpt", "value": "visual quality: 2\n temporal consistency: 1\n dynamic degree: 3\n text-to-video alignment: 2\n factual consistency: 2\n" } ]

[ "7000513_00.jpg", "7000513_01.jpg", "7000513_02.jpg", "7000513_03.jpg", "7000513_04.jpg", "7000513_05.jpg", "7000513_06.jpg", "7000513_07.jpg" ]

[ { "from": "human", "value": "\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos,\nplease watch the following frames of a given video and see the text prompt for generating the video,\nthen give scores from 5 different dimensions:\n(1) visual quality: the quality of the video in terms of clearness, resolution, brightness, and color\n(2) temporal consistency, both the consistency of objects or humans and the smoothness of motion or movements\n(3) dynamic degree, the degree of dynamic changes\n(4) text-to-video alignment, the alignment between the text prompt and the video content\n(5) factual consistency, the consistency of the video content with the common-sense and factual knowledge\n\nfor each dimension, output a float number from 1.0 to 4.0,\nthe higher the number is, the better the video performs in that sub-score, \nthe lowest 1.0 means Bad, the highest 4.0 means Perfect/Real (the video is like a real video)\nHere is an output example:\nvisual quality: 3.2\ntemporal consistency: 2.7\ndynamic degree: 4.0\ntext-to-video alignment: 2.3\nfactual consistency: 1.8\n\nFor this video, the text prompt is \"Cells that eat amoebas, which in turn eat other amoebas.\",\nall the frames of video are as follows:\n\n<image> <image> <image> <image> <image> <image> <image> <image> " }, { "from": "gpt", "value": "visual quality: 3\n temporal consistency: 3\n dynamic degree: 3\n text-to-video alignment: 2\n factual consistency: 3\n" } ]

[ "2002909_00.jpg", "2002909_01.jpg", "2002909_02.jpg", "2002909_03.jpg", "2002909_04.jpg", "2002909_05.jpg", "2002909_06.jpg", "2002909_07.jpg", "2002909_08.jpg", "2002909_09.jpg", "2002909_10.jpg", "2002909_11.jpg", "2002909_12.jpg", "2002909_13.jpg", "2002909_14.jpg", "2002909_15.jpg" ]

[ { "from": "human", "value": "\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos,\nplease watch the following frames of a given video and see the text prompt for generating the video,\nthen give scores from 5 different dimensions:\n(1) visual quality: the quality of the video in terms of clearness, resolution, brightness, and color\n(2) temporal consistency, both the consistency of objects or humans and the smoothness of motion or movements\n(3) dynamic degree, the degree of dynamic changes\n(4) text-to-video alignment, the alignment between the text prompt and the video content\n(5) factual consistency, the consistency of the video content with the common-sense and factual knowledge\n\nfor each dimension, output a float number from 1.0 to 4.0,\nthe higher the number is, the better the video performs in that sub-score, \nthe lowest 1.0 means Bad, the highest 4.0 means Perfect/Real (the video is like a real video)\nHere is an output example:\nvisual quality: 3.2\ntemporal consistency: 2.7\ndynamic degree: 4.0\ntext-to-video alignment: 2.3\nfactual consistency: 1.8\n\nFor this video, the text prompt is \"Detail the bedtime rituals, such as patting the cows, whispering sweet words to the sheep, and giving the chickens a gentle lullaby. 3d image animation gs 20 motion 4\",\nall the frames of video are as follows:\n\n<image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> " }, { "from": "gpt", "value": "visual quality: 2\n temporal consistency: 3\n dynamic degree: 3\n text-to-video alignment: 1\n factual consistency: 3\n" } ]

[ "3000222_00.jpg", "3000222_01.jpg", "3000222_02.jpg", "3000222_03.jpg", "3000222_04.jpg", "3000222_05.jpg", "3000222_06.jpg", "3000222_07.jpg", "3000222_08.jpg", "3000222_09.jpg", "3000222_10.jpg", "3000222_11.jpg", "3000222_12.jpg", "3000222_13.jpg", "3000222_14.jpg", "3000222_15.jpg" ]

[ { "from": "human", "value": "\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos,\nplease watch the following frames of a given video and see the text prompt for generating the video,\nthen give scores from 5 different dimensions:\n(1) visual quality: the quality of the video in terms of clearness, resolution, brightness, and color\n(2) temporal consistency, both the consistency of objects or humans and the smoothness of motion or movements\n(3) dynamic degree, the degree of dynamic changes\n(4) text-to-video alignment, the alignment between the text prompt and the video content\n(5) factual consistency, the consistency of the video content with the common-sense and factual knowledge\n\nfor each dimension, output a float number from 1.0 to 4.0,\nthe higher the number is, the better the video performs in that sub-score, \nthe lowest 1.0 means Bad, the highest 4.0 means Perfect/Real (the video is like a real video)\nHere is an output example:\nvisual quality: 3.2\ntemporal consistency: 2.7\ndynamic degree: 4.0\ntext-to-video alignment: 2.3\nfactual consistency: 1.8\n\nFor this video, the text prompt is \"portal to heaven opening in the night sky\",\nall the frames of video are as follows:\n\n<image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> " }, { "from": "gpt", "value": "visual quality: 2\n temporal consistency: 2\n dynamic degree: 3\n text-to-video alignment: 1\n factual consistency: 1\n" } ]

[ "0002382_00.jpg", "0002382_01.jpg", "0002382_02.jpg", "0002382_03.jpg", "0002382_04.jpg", "0002382_05.jpg", "0002382_06.jpg", "0002382_07.jpg", "0002382_08.jpg", "0002382_09.jpg", "0002382_10.jpg", "0002382_11.jpg", "0002382_12.jpg", "0002382_13.jpg", "0002382_14.jpg", "0002382_15.jpg", "0002382_16.jpg", "0002382_17.jpg", "0002382_18.jpg", "0002382_19.jpg", "0002382_20.jpg", "0002382_21.jpg", "0002382_22.jpg", "0002382_23.jpg" ]

[ { "from": "human", "value": "\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos,\nplease watch the following frames of a given video and see the text prompt for generating the video,\nthen give scores from 5 different dimensions:\n(1) visual quality: the quality of the video in terms of clearness, resolution, brightness, and color\n(2) temporal consistency, both the consistency of objects or humans and the smoothness of motion or movements\n(3) dynamic degree, the degree of dynamic changes\n(4) text-to-video alignment, the alignment between the text prompt and the video content\n(5) factual consistency, the consistency of the video content with the common-sense and factual knowledge\n\nfor each dimension, output a float number from 1.0 to 4.0,\nthe higher the number is, the better the video performs in that sub-score, \nthe lowest 1.0 means Bad, the highest 4.0 means Perfect/Real (the video is like a real video)\nHere is an output example:\nvisual quality: 3.2\ntemporal consistency: 2.7\ndynamic degree: 4.0\ntext-to-video alignment: 2.3\nfactual consistency: 1.8\n\nFor this video, the text prompt is \"A luxurious Car Garage contaning Sports Cars and Lowriders\",\nall the frames of video are as follows:\n\n<image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> " }, { "from": "gpt", "value": "visual quality: 3\n temporal consistency: 3\n dynamic degree: 1\n text-to-video alignment: 3\n factual consistency: 3\n" } ]

[ "3001307_00.jpg", "3001307_01.jpg", "3001307_02.jpg", "3001307_03.jpg", "3001307_04.jpg", "3001307_05.jpg", "3001307_06.jpg", "3001307_07.jpg", "3001307_08.jpg", "3001307_09.jpg", "3001307_10.jpg", "3001307_11.jpg", "3001307_12.jpg", "3001307_13.jpg", "3001307_14.jpg", "3001307_15.jpg" ]

[ { "from": "human", "value": "\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos,\nplease watch the following frames of a given video and see the text prompt for generating the video,\nthen give scores from 5 different dimensions:\n(1) visual quality: the quality of the video in terms of clearness, resolution, brightness, and color\n(2) temporal consistency, both the consistency of objects or humans and the smoothness of motion or movements\n(3) dynamic degree, the degree of dynamic changes\n(4) text-to-video alignment, the alignment between the text prompt and the video content\n(5) factual consistency, the consistency of the video content with the common-sense and factual knowledge\n\nfor each dimension, output a float number from 1.0 to 4.0,\nthe higher the number is, the better the video performs in that sub-score, \nthe lowest 1.0 means Bad, the highest 4.0 means Perfect/Real (the video is like a real video)\nHere is an output example:\nvisual quality: 3.2\ntemporal consistency: 2.7\ndynamic degree: 4.0\ntext-to-video alignment: 2.3\nfactual consistency: 1.8\n\nFor this video, the text prompt is \"The cat runs away from the mice\",\nall the frames of video are as follows:\n\n<image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> " }, { "from": "gpt", "value": "visual quality: 3\n temporal consistency: 2\n dynamic degree: 3\n text-to-video alignment: 2\n factual consistency: 3\n" } ]

[ "3002396_00.jpg", "3002396_01.jpg", "3002396_02.jpg", "3002396_03.jpg", "3002396_04.jpg", "3002396_05.jpg", "3002396_06.jpg", "3002396_07.jpg", "3002396_08.jpg", "3002396_09.jpg", "3002396_10.jpg", "3002396_11.jpg", "3002396_12.jpg", "3002396_13.jpg", "3002396_14.jpg", "3002396_15.jpg" ]

[ { "from": "human", "value": "\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos,\nplease watch the following frames of a given video and see the text prompt for generating the video,\nthen give scores from 5 different dimensions:\n(1) visual quality: the quality of the video in terms of clearness, resolution, brightness, and color\n(2) temporal consistency, both the consistency of objects or humans and the smoothness of motion or movements\n(3) dynamic degree, the degree of dynamic changes\n(4) text-to-video alignment, the alignment between the text prompt and the video content\n(5) factual consistency, the consistency of the video content with the common-sense and factual knowledge\n\nfor each dimension, output a float number from 1.0 to 4.0,\nthe higher the number is, the better the video performs in that sub-score, \nthe lowest 1.0 means Bad, the highest 4.0 means Perfect/Real (the video is like a real video)\nHere is an output example:\nvisual quality: 3.2\ntemporal consistency: 2.7\ndynamic degree: 4.0\ntext-to-video alignment: 2.3\nfactual consistency: 1.8\n\nFor this video, the text prompt is \"point of view video of someone falling from a tall building\",\nall the frames of video are as follows:\n\n<image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> " }, { "from": "gpt", "value": "visual quality: 2\n temporal consistency: 1\n dynamic degree: 3\n text-to-video alignment: 1\n factual consistency: 1\n" } ]

[ "0002230_00.jpg", "0002230_01.jpg", "0002230_02.jpg", "0002230_03.jpg", "0002230_04.jpg", "0002230_05.jpg", "0002230_06.jpg", "0002230_07.jpg", "0002230_08.jpg", "0002230_09.jpg", "0002230_10.jpg", "0002230_11.jpg", "0002230_12.jpg", "0002230_13.jpg", "0002230_14.jpg", "0002230_15.jpg", "0002230_16.jpg", "0002230_17.jpg", "0002230_18.jpg", "0002230_19.jpg", "0002230_20.jpg", "0002230_21.jpg", "0002230_22.jpg", "0002230_23.jpg" ]

[ { "from": "human", "value": "\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos,\nplease watch the following frames of a given video and see the text prompt for generating the video,\nthen give scores from 5 different dimensions:\n(1) visual quality: the quality of the video in terms of clearness, resolution, brightness, and color\n(2) temporal consistency, both the consistency of objects or humans and the smoothness of motion or movements\n(3) dynamic degree, the degree of dynamic changes\n(4) text-to-video alignment, the alignment between the text prompt and the video content\n(5) factual consistency, the consistency of the video content with the common-sense and factual knowledge\n\nfor each dimension, output a float number from 1.0 to 4.0,\nthe higher the number is, the better the video performs in that sub-score, \nthe lowest 1.0 means Bad, the highest 4.0 means Perfect/Real (the video is like a real video)\nHere is an output example:\nvisual quality: 3.2\ntemporal consistency: 2.7\ndynamic degree: 4.0\ntext-to-video alignment: 2.3\nfactual consistency: 1.8\n\nFor this video, the text prompt is \"a people rise a horse on the playground\",\nall the frames of video are as follows:\n\n<image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> " }, { "from": "gpt", "value": "visual quality: 3\n temporal consistency: 3\n dynamic degree: 2\n text-to-video alignment: 1\n factual consistency: 3\n" } ]

[ "7000727_00.jpg", "7000727_01.jpg", "7000727_02.jpg", "7000727_03.jpg", "7000727_04.jpg", "7000727_05.jpg", "7000727_06.jpg", "7000727_07.jpg" ]

[ { "from": "human", "value": "\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos,\nplease watch the following frames of a given video and see the text prompt for generating the video,\nthen give scores from 5 different dimensions:\n(1) visual quality: the quality of the video in terms of clearness, resolution, brightness, and color\n(2) temporal consistency, both the consistency of objects or humans and the smoothness of motion or movements\n(3) dynamic degree, the degree of dynamic changes\n(4) text-to-video alignment, the alignment between the text prompt and the video content\n(5) factual consistency, the consistency of the video content with the common-sense and factual knowledge\n\nfor each dimension, output a float number from 1.0 to 4.0,\nthe higher the number is, the better the video performs in that sub-score, \nthe lowest 1.0 means Bad, the highest 4.0 means Perfect/Real (the video is like a real video)\nHere is an output example:\nvisual quality: 3.2\ntemporal consistency: 2.7\ndynamic degree: 4.0\ntext-to-video alignment: 2.3\nfactual consistency: 1.8\n\nFor this video, the text prompt is \"Imagine your world\n no beer to drink\n imagine your life\n without excesses of truth\n imagine the courtship\n no premarital marriage\n imagine that this\n was reality\",\nall the frames of video are as follows:\n\n<image> <image> <image> <image> <image> <image> <image> <image> " }, { "from": "gpt", "value": "visual quality: 2\n temporal consistency: 3\n dynamic degree: 3\n text-to-video alignment: 2\n factual consistency: 3\n" } ]

[ "1001791_00.jpg", "1001791_01.jpg", "1001791_02.jpg", "1001791_03.jpg", "1001791_04.jpg", "1001791_05.jpg", "1001791_06.jpg", "1001791_07.jpg", "1001791_08.jpg", "1001791_09.jpg", "1001791_10.jpg", "1001791_11.jpg", "1001791_12.jpg", "1001791_13.jpg", "1001791_14.jpg", "1001791_15.jpg" ]

[ { "from": "human", "value": "\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos,\nplease watch the following frames of a given video and see the text prompt for generating the video,\nthen give scores from 5 different dimensions:\n(1) visual quality: the quality of the video in terms of clearness, resolution, brightness, and color\n(2) temporal consistency, both the consistency of objects or humans and the smoothness of motion or movements\n(3) dynamic degree, the degree of dynamic changes\n(4) text-to-video alignment, the alignment between the text prompt and the video content\n(5) factual consistency, the consistency of the video content with the common-sense and factual knowledge\n\nfor each dimension, output a float number from 1.0 to 4.0,\nthe higher the number is, the better the video performs in that sub-score, \nthe lowest 1.0 means Bad, the highest 4.0 means Perfect/Real (the video is like a real video)\nHere is an output example:\nvisual quality: 3.2\ntemporal consistency: 2.7\ndynamic degree: 4.0\ntext-to-video alignment: 2.3\nfactual consistency: 1.8\n\nFor this video, the text prompt is \"detailed image of old man drinking a cup of coffee\",\nall the frames of video are as follows:\n\n<image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> " }, { "from": "gpt", "value": "visual quality: 1\n temporal consistency: 1\n dynamic degree: 3\n text-to-video alignment: 2\n factual consistency: 1\n" } ]

[ "3000389_00.jpg", "3000389_01.jpg", "3000389_02.jpg", "3000389_03.jpg", "3000389_04.jpg", "3000389_05.jpg", "3000389_06.jpg", "3000389_07.jpg", "3000389_08.jpg", "3000389_09.jpg", "3000389_10.jpg", "3000389_11.jpg", "3000389_12.jpg", "3000389_13.jpg", "3000389_14.jpg", "3000389_15.jpg" ]

[ { "from": "human", "value": "\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos,\nplease watch the following frames of a given video and see the text prompt for generating the video,\nthen give scores from 5 different dimensions:\n(1) visual quality: the quality of the video in terms of clearness, resolution, brightness, and color\n(2) temporal consistency, both the consistency of objects or humans and the smoothness of motion or movements\n(3) dynamic degree, the degree of dynamic changes\n(4) text-to-video alignment, the alignment between the text prompt and the video content\n(5) factual consistency, the consistency of the video content with the common-sense and factual knowledge\n\nfor each dimension, output a float number from 1.0 to 4.0,\nthe higher the number is, the better the video performs in that sub-score, \nthe lowest 1.0 means Bad, the highest 4.0 means Perfect/Real (the video is like a real video)\nHere is an output example:\nvisual quality: 3.2\ntemporal consistency: 2.7\ndynamic degree: 4.0\ntext-to-video alignment: 2.3\nfactual consistency: 1.8\n\nFor this video, the text prompt is \"two silhouette of man, fighting each other, bare hand, martial arts, purple otherworldly aura emanating from body\",\nall the frames of video are as follows:\n\n<image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> " }, { "from": "gpt", "value": "visual quality: 2\n temporal consistency: 1\n dynamic degree: 3\n text-to-video alignment: 1\n factual consistency: 1\n" } ]

[ "0000602_00.jpg", "0000602_01.jpg", "0000602_02.jpg", "0000602_03.jpg", "0000602_04.jpg", "0000602_05.jpg", "0000602_06.jpg", "0000602_07.jpg", "0000602_08.jpg", "0000602_09.jpg", "0000602_10.jpg", "0000602_11.jpg", "0000602_12.jpg", "0000602_13.jpg", "0000602_14.jpg", "0000602_15.jpg", "0000602_16.jpg", "0000602_17.jpg", "0000602_18.jpg", "0000602_19.jpg", "0000602_20.jpg", "0000602_21.jpg", "0000602_22.jpg", "0000602_23.jpg" ]

[ { "from": "human", "value": "\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos,\nplease watch the following frames of a given video and see the text prompt for generating the video,\nthen give scores from 5 different dimensions:\n(1) visual quality: the quality of the video in terms of clearness, resolution, brightness, and color\n(2) temporal consistency, both the consistency of objects or humans and the smoothness of motion or movements\n(3) dynamic degree, the degree of dynamic changes\n(4) text-to-video alignment, the alignment between the text prompt and the video content\n(5) factual consistency, the consistency of the video content with the common-sense and factual knowledge\n\nfor each dimension, output a float number from 1.0 to 4.0,\nthe higher the number is, the better the video performs in that sub-score, \nthe lowest 1.0 means Bad, the highest 4.0 means Perfect/Real (the video is like a real video)\nHere is an output example:\nvisual quality: 3.2\ntemporal consistency: 2.7\ndynamic degree: 4.0\ntext-to-video alignment: 2.3\nfactual consistency: 1.8\n\nFor this video, the text prompt is \"young girl cartoon style. looks like Rapunzel. 3д. 25 years old. brown straight hair. beautiful. eyes closed. she is sleeping. her head is on the pillow. the girl is wearing pajamas. the room is dark.\",\nall the frames of video are as follows:\n\n<image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> " }, { "from": "gpt", "value": "visual quality: 3\n temporal consistency: 3\n dynamic degree: 1\n text-to-video alignment: 3\n factual consistency: 3\n" } ]

[ "5000291_00.jpg", "5000291_01.jpg", "5000291_02.jpg", "5000291_03.jpg", "5000291_04.jpg", "5000291_05.jpg", "5000291_06.jpg", "5000291_07.jpg", "5000291_08.jpg", "5000291_09.jpg", "5000291_10.jpg", "5000291_11.jpg", "5000291_12.jpg", "5000291_13.jpg", "5000291_14.jpg", "5000291_15.jpg" ]

[ { "from": "human", "value": "\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos,\nplease watch the following frames of a given video and see the text prompt for generating the video,\nthen give scores from 5 different dimensions:\n(1) visual quality: the quality of the video in terms of clearness, resolution, brightness, and color\n(2) temporal consistency, both the consistency of objects or humans and the smoothness of motion or movements\n(3) dynamic degree, the degree of dynamic changes\n(4) text-to-video alignment, the alignment between the text prompt and the video content\n(5) factual consistency, the consistency of the video content with the common-sense and factual knowledge\n\nfor each dimension, output a float number from 1.0 to 4.0,\nthe higher the number is, the better the video performs in that sub-score, \nthe lowest 1.0 means Bad, the highest 4.0 means Perfect/Real (the video is like a real video)\nHere is an output example:\nvisual quality: 3.2\ntemporal consistency: 2.7\ndynamic degree: 4.0\ntext-to-video alignment: 2.3\nfactual consistency: 1.8\n\nFor this video, the text prompt is \"The camera pans gently through a dense green forest, huge trees stretching towards the sky, and sunlight creeping through the branches, Lord of the Rings movie style\",\nall the frames of video are as follows:\n\n<image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> " }, { "from": "gpt", "value": "visual quality: 3\n temporal consistency: 3\n dynamic degree: 3\n text-to-video alignment: 3\n factual consistency: 3\n" } ]

[ "7000280_00.jpg", "7000280_01.jpg", "7000280_02.jpg", "7000280_03.jpg", "7000280_04.jpg", "7000280_05.jpg", "7000280_06.jpg", "7000280_07.jpg" ]

[ { "from": "human", "value": "\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos,\nplease watch the following frames of a given video and see the text prompt for generating the video,\nthen give scores from 5 different dimensions:\n(1) visual quality: the quality of the video in terms of clearness, resolution, brightness, and color\n(2) temporal consistency, both the consistency of objects or humans and the smoothness of motion or movements\n(3) dynamic degree, the degree of dynamic changes\n(4) text-to-video alignment, the alignment between the text prompt and the video content\n(5) factual consistency, the consistency of the video content with the common-sense and factual knowledge\n\nfor each dimension, output a float number from 1.0 to 4.0,\nthe higher the number is, the better the video performs in that sub-score, \nthe lowest 1.0 means Bad, the highest 4.0 means Perfect/Real (the video is like a real video)\nHere is an output example:\nvisual quality: 3.2\ntemporal consistency: 2.7\ndynamic degree: 4.0\ntext-to-video alignment: 2.3\nfactual consistency: 1.8\n\nFor this video, the text prompt is \"guitar player running in a flower field smiling, cinematic style\",\nall the frames of video are as follows:\n\n<image> <image> <image> <image> <image> <image> <image> <image> " }, { "from": "gpt", "value": "visual quality: 2\n temporal consistency: 2\n dynamic degree: 3\n text-to-video alignment: 3\n factual consistency: 2\n" } ]

[ "6000571_00.jpg", "6000571_01.jpg", "6000571_02.jpg", "6000571_03.jpg", "6000571_04.jpg", "6000571_05.jpg", "6000571_06.jpg", "6000571_07.jpg", "6000571_08.jpg", "6000571_09.jpg", "6000571_10.jpg", "6000571_11.jpg", "6000571_12.jpg", "6000571_13.jpg", "6000571_14.jpg", "6000571_15.jpg" ]

[ { "from": "human", "value": "\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos,\nplease watch the following frames of a given video and see the text prompt for generating the video,\nthen give scores from 5 different dimensions:\n(1) visual quality: the quality of the video in terms of clearness, resolution, brightness, and color\n(2) temporal consistency, both the consistency of objects or humans and the smoothness of motion or movements\n(3) dynamic degree, the degree of dynamic changes\n(4) text-to-video alignment, the alignment between the text prompt and the video content\n(5) factual consistency, the consistency of the video content with the common-sense and factual knowledge\n\nfor each dimension, output a float number from 1.0 to 4.0,\nthe higher the number is, the better the video performs in that sub-score, \nthe lowest 1.0 means Bad, the highest 4.0 means Perfect/Real (the video is like a real video)\nHere is an output example:\nvisual quality: 3.2\ntemporal consistency: 2.7\ndynamic degree: 4.0\ntext-to-video alignment: 2.3\nfactual consistency: 1.8\n\nFor this video, the text prompt is \"a video in which Jesus meets a blind person.\",\nall the frames of video are as follows:\n\n<image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> " }, { "from": "gpt", "value": "visual quality: 1\n temporal consistency: 1\n dynamic degree: 3\n text-to-video alignment: 3\n factual consistency: 1\n" } ]

[ "6000740_00.jpg", "6000740_01.jpg", "6000740_02.jpg", "6000740_03.jpg", "6000740_04.jpg", "6000740_05.jpg", "6000740_06.jpg", "6000740_07.jpg", "6000740_08.jpg", "6000740_09.jpg", "6000740_10.jpg", "6000740_11.jpg", "6000740_12.jpg", "6000740_13.jpg", "6000740_14.jpg", "6000740_15.jpg" ]

[ { "from": "human", "value": "\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos,\nplease watch the following frames of a given video and see the text prompt for generating the video,\nthen give scores from 5 different dimensions:\n(1) visual quality: the quality of the video in terms of clearness, resolution, brightness, and color\n(2) temporal consistency, both the consistency of objects or humans and the smoothness of motion or movements\n(3) dynamic degree, the degree of dynamic changes\n(4) text-to-video alignment, the alignment between the text prompt and the video content\n(5) factual consistency, the consistency of the video content with the common-sense and factual knowledge\n\nfor each dimension, output a float number from 1.0 to 4.0,\nthe higher the number is, the better the video performs in that sub-score, \nthe lowest 1.0 means Bad, the highest 4.0 means Perfect/Real (the video is like a real video)\nHere is an output example:\nvisual quality: 3.2\ntemporal consistency: 2.7\ndynamic degree: 4.0\ntext-to-video alignment: 2.3\nfactual consistency: 1.8\n\nFor this video, the text prompt is \"A large caravan consisting of camels, and above them are people wearing Bedouin clothing preparing to leave in the morning on the camel. A veiled woman sits inside a howdah. She raises the curtain to see in front of her.\",\nall the frames of video are as follows:\n\n<image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> " }, { "from": "gpt", "value": "visual quality: 1\n temporal consistency: 2\n dynamic degree: 3\n text-to-video alignment: 1\n factual consistency: 1\n" } ]

[ "1002072_00.jpg", "1002072_01.jpg", "1002072_02.jpg", "1002072_03.jpg", "1002072_04.jpg", "1002072_05.jpg", "1002072_06.jpg", "1002072_07.jpg", "1002072_08.jpg", "1002072_09.jpg", "1002072_10.jpg", "1002072_11.jpg", "1002072_12.jpg", "1002072_13.jpg", "1002072_14.jpg", "1002072_15.jpg" ]

[ { "from": "human", "value": "\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos,\nplease watch the following frames of a given video and see the text prompt for generating the video,\nthen give scores from 5 different dimensions:\n(1) visual quality: the quality of the video in terms of clearness, resolution, brightness, and color\n(2) temporal consistency, both the consistency of objects or humans and the smoothness of motion or movements\n(3) dynamic degree, the degree of dynamic changes\n(4) text-to-video alignment, the alignment between the text prompt and the video content\n(5) factual consistency, the consistency of the video content with the common-sense and factual knowledge\n\nfor each dimension, output a float number from 1.0 to 4.0,\nthe higher the number is, the better the video performs in that sub-score, \nthe lowest 1.0 means Bad, the highest 4.0 means Perfect/Real (the video is like a real video)\nHere is an output example:\nvisual quality: 3.2\ntemporal consistency: 2.7\ndynamic degree: 4.0\ntext-to-video alignment: 2.3\nfactual consistency: 1.8\n\nFor this video, the text prompt is \"A hole in side the tree\",\nall the frames of video are as follows:\n\n<image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> " }, { "from": "gpt", "value": "visual quality: 1\n temporal consistency: 2\n dynamic degree: 3\n text-to-video alignment: 2\n factual consistency: 1\n" } ]

[ "4001038_00.jpg", "4001038_01.jpg", "4001038_02.jpg", "4001038_03.jpg", "4001038_04.jpg", "4001038_05.jpg", "4001038_06.jpg", "4001038_07.jpg", "4001038_08.jpg", "4001038_09.jpg", "4001038_10.jpg", "4001038_11.jpg", "4001038_12.jpg", "4001038_13.jpg", "4001038_14.jpg", "4001038_15.jpg" ]

[ { "from": "human", "value": "\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos,\nplease watch the following frames of a given video and see the text prompt for generating the video,\nthen give scores from 5 different dimensions:\n(1) visual quality: the quality of the video in terms of clearness, resolution, brightness, and color\n(2) temporal consistency, both the consistency of objects or humans and the smoothness of motion or movements\n(3) dynamic degree, the degree of dynamic changes\n(4) text-to-video alignment, the alignment between the text prompt and the video content\n(5) factual consistency, the consistency of the video content with the common-sense and factual knowledge\n\nfor each dimension, output a float number from 1.0 to 4.0,\nthe higher the number is, the better the video performs in that sub-score, \nthe lowest 1.0 means Bad, the highest 4.0 means Perfect/Real (the video is like a real video)\nHere is an output example:\nvisual quality: 3.2\ntemporal consistency: 2.7\ndynamic degree: 4.0\ntext-to-video alignment: 2.3\nfactual consistency: 1.8\n\nFor this video, the text prompt is \"A Ninja creating a Storm by Rotating in the Same Place\",\nall the frames of video are as follows:\n\n<image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> " }, { "from": "gpt", "value": "visual quality: 3\n temporal consistency: 2\n dynamic degree: 3\n text-to-video alignment: 2\n factual consistency: 3\n" } ]

[ "0001518_00.jpg", "0001518_01.jpg", "0001518_02.jpg", "0001518_03.jpg", "0001518_04.jpg", "0001518_05.jpg", "0001518_06.jpg", "0001518_07.jpg", "0001518_08.jpg", "0001518_09.jpg", "0001518_10.jpg", "0001518_11.jpg", "0001518_12.jpg", "0001518_13.jpg", "0001518_14.jpg", "0001518_15.jpg", "0001518_16.jpg", "0001518_17.jpg", "0001518_18.jpg", "0001518_19.jpg", "0001518_20.jpg", "0001518_21.jpg", "0001518_22.jpg", "0001518_23.jpg" ]

[ { "from": "human", "value": "\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos,\nplease watch the following frames of a given video and see the text prompt for generating the video,\nthen give scores from 5 different dimensions:\n(1) visual quality: the quality of the video in terms of clearness, resolution, brightness, and color\n(2) temporal consistency, both the consistency of objects or humans and the smoothness of motion or movements\n(3) dynamic degree, the degree of dynamic changes\n(4) text-to-video alignment, the alignment between the text prompt and the video content\n(5) factual consistency, the consistency of the video content with the common-sense and factual knowledge\n\nfor each dimension, output a float number from 1.0 to 4.0,\nthe higher the number is, the better the video performs in that sub-score, \nthe lowest 1.0 means Bad, the highest 4.0 means Perfect/Real (the video is like a real video)\nHere is an output example:\nvisual quality: 3.2\ntemporal consistency: 2.7\ndynamic degree: 4.0\ntext-to-video alignment: 2.3\nfactual consistency: 1.8\n\nFor this video, the text prompt is \"Tesla Model S, flying with wings, on the sea, there are several islands in the sea, and there are dinosaurs swimming in the sea\",\nall the frames of video are as follows:\n\n<image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> " }, { "from": "gpt", "value": "visual quality: 3\n temporal consistency: 3\n dynamic degree: 2\n text-to-video alignment: 3\n factual consistency: 2\n" } ]

[ "7000205_00.jpg", "7000205_01.jpg", "7000205_02.jpg", "7000205_03.jpg", "7000205_04.jpg", "7000205_05.jpg", "7000205_06.jpg", "7000205_07.jpg" ]

[ { "from": "human", "value": "\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos,\nplease watch the following frames of a given video and see the text prompt for generating the video,\nthen give scores from 5 different dimensions:\n(1) visual quality: the quality of the video in terms of clearness, resolution, brightness, and color\n(2) temporal consistency, both the consistency of objects or humans and the smoothness of motion or movements\n(3) dynamic degree, the degree of dynamic changes\n(4) text-to-video alignment, the alignment between the text prompt and the video content\n(5) factual consistency, the consistency of the video content with the common-sense and factual knowledge\n\nfor each dimension, output a float number from 1.0 to 4.0,\nthe higher the number is, the better the video performs in that sub-score, \nthe lowest 1.0 means Bad, the highest 4.0 means Perfect/Real (the video is like a real video)\nHere is an output example:\nvisual quality: 3.2\ntemporal consistency: 2.7\ndynamic degree: 4.0\ntext-to-video alignment: 2.3\nfactual consistency: 1.8\n\nFor this video, the text prompt is \"baby goat hiding in the bush\",\nall the frames of video are as follows:\n\n<image> <image> <image> <image> <image> <image> <image> <image> " }, { "from": "gpt", "value": "visual quality: 2\n temporal consistency: 2\n dynamic degree: 3\n text-to-video alignment: 3\n factual consistency: 3\n" } ]

[ "3000881_00.jpg", "3000881_01.jpg", "3000881_02.jpg", "3000881_03.jpg", "3000881_04.jpg", "3000881_05.jpg", "3000881_06.jpg", "3000881_07.jpg", "3000881_08.jpg", "3000881_09.jpg", "3000881_10.jpg", "3000881_11.jpg", "3000881_12.jpg", "3000881_13.jpg", "3000881_14.jpg", "3000881_15.jpg" ]

[ { "from": "human", "value": "\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos,\nplease watch the following frames of a given video and see the text prompt for generating the video,\nthen give scores from 5 different dimensions:\n(1) visual quality: the quality of the video in terms of clearness, resolution, brightness, and color\n(2) temporal consistency, both the consistency of objects or humans and the smoothness of motion or movements\n(3) dynamic degree, the degree of dynamic changes\n(4) text-to-video alignment, the alignment between the text prompt and the video content\n(5) factual consistency, the consistency of the video content with the common-sense and factual knowledge\n\nfor each dimension, output a float number from 1.0 to 4.0,\nthe higher the number is, the better the video performs in that sub-score, \nthe lowest 1.0 means Bad, the highest 4.0 means Perfect/Real (the video is like a real video)\nHere is an output example:\nvisual quality: 3.2\ntemporal consistency: 2.7\ndynamic degree: 4.0\ntext-to-video alignment: 2.3\nfactual consistency: 1.8\n\nFor this video, the text prompt is \"tracking shot of an pink astronaut in space\",\nall the frames of video are as follows:\n\n<image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> " }, { "from": "gpt", "value": "visual quality: 2\n temporal consistency: 2\n dynamic degree: 3\n text-to-video alignment: 3\n factual consistency: 2\n" } ]

[ "0002991_00.jpg", "0002991_01.jpg", "0002991_02.jpg", "0002991_03.jpg", "0002991_04.jpg", "0002991_05.jpg", "0002991_06.jpg", "0002991_07.jpg", "0002991_08.jpg", "0002991_09.jpg", "0002991_10.jpg", "0002991_11.jpg", "0002991_12.jpg", "0002991_13.jpg", "0002991_14.jpg", "0002991_15.jpg", "0002991_16.jpg", "0002991_17.jpg", "0002991_18.jpg", "0002991_19.jpg", "0002991_20.jpg", "0002991_21.jpg", "0002991_22.jpg", "0002991_23.jpg" ]

[ { "from": "human", "value": "\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos,\nplease watch the following frames of a given video and see the text prompt for generating the video,\nthen give scores from 5 different dimensions:\n(1) visual quality: the quality of the video in terms of clearness, resolution, brightness, and color\n(2) temporal consistency, both the consistency of objects or humans and the smoothness of motion or movements\n(3) dynamic degree, the degree of dynamic changes\n(4) text-to-video alignment, the alignment between the text prompt and the video content\n(5) factual consistency, the consistency of the video content with the common-sense and factual knowledge\n\nfor each dimension, output a float number from 1.0 to 4.0,\nthe higher the number is, the better the video performs in that sub-score, \nthe lowest 1.0 means Bad, the highest 4.0 means Perfect/Real (the video is like a real video)\nHere is an output example:\nvisual quality: 3.2\ntemporal consistency: 2.7\ndynamic degree: 4.0\ntext-to-video alignment: 2.3\nfactual consistency: 1.8\n\nFor this video, the text prompt is \"A joyful and energetic person walking along an endless beach in the sunset while a melodic whistling sound of joy plays in the background, the sun reflecting off the water like a river of gold in a tropical paradise.\",\nall the frames of video are as follows:\n\n<image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> " }, { "from": "gpt", "value": "visual quality: 2\n temporal consistency: 3\n dynamic degree: 2\n text-to-video alignment: 3\n factual consistency: 3\n" } ]

[ "5000785_00.jpg", "5000785_01.jpg", "5000785_02.jpg", "5000785_03.jpg", "5000785_04.jpg", "5000785_05.jpg", "5000785_06.jpg", "5000785_07.jpg", "5000785_08.jpg", "5000785_09.jpg", "5000785_10.jpg", "5000785_11.jpg", "5000785_12.jpg", "5000785_13.jpg", "5000785_14.jpg", "5000785_15.jpg" ]

[ { "from": "human", "value": "\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos,\nplease watch the following frames of a given video and see the text prompt for generating the video,\nthen give scores from 5 different dimensions:\n(1) visual quality: the quality of the video in terms of clearness, resolution, brightness, and color\n(2) temporal consistency, both the consistency of objects or humans and the smoothness of motion or movements\n(3) dynamic degree, the degree of dynamic changes\n(4) text-to-video alignment, the alignment between the text prompt and the video content\n(5) factual consistency, the consistency of the video content with the common-sense and factual knowledge\n\nfor each dimension, output a float number from 1.0 to 4.0,\nthe higher the number is, the better the video performs in that sub-score, \nthe lowest 1.0 means Bad, the highest 4.0 means Perfect/Real (the video is like a real video)\nHere is an output example:\nvisual quality: 3.2\ntemporal consistency: 2.7\ndynamic degree: 4.0\ntext-to-video alignment: 2.3\nfactual consistency: 1.8\n\nFor this video, the text prompt is \"zoom in an artistic collage of Chinese Dragon head ，8k，masterpiece\",\nall the frames of video are as follows:\n\n<image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> " }, { "from": "gpt", "value": "visual quality: 3\n temporal consistency: 3\n dynamic degree: 2\n text-to-video alignment: 3\n factual consistency: 1\n" } ]

[ "0002305_00.jpg", "0002305_01.jpg", "0002305_02.jpg", "0002305_03.jpg", "0002305_04.jpg", "0002305_05.jpg", "0002305_06.jpg", "0002305_07.jpg", "0002305_08.jpg", "0002305_09.jpg", "0002305_10.jpg", "0002305_11.jpg", "0002305_12.jpg", "0002305_13.jpg", "0002305_14.jpg", "0002305_15.jpg", "0002305_16.jpg", "0002305_17.jpg", "0002305_18.jpg", "0002305_19.jpg", "0002305_20.jpg", "0002305_21.jpg", "0002305_22.jpg", "0002305_23.jpg" ]

[ { "from": "human", "value": "\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos,\nplease watch the following frames of a given video and see the text prompt for generating the video,\nthen give scores from 5 different dimensions:\n(1) visual quality: the quality of the video in terms of clearness, resolution, brightness, and color\n(2) temporal consistency, both the consistency of objects or humans and the smoothness of motion or movements\n(3) dynamic degree, the degree of dynamic changes\n(4) text-to-video alignment, the alignment between the text prompt and the video content\n(5) factual consistency, the consistency of the video content with the common-sense and factual knowledge\n\nfor each dimension, output a float number from 1.0 to 4.0,\nthe higher the number is, the better the video performs in that sub-score, \nthe lowest 1.0 means Bad, the highest 4.0 means Perfect/Real (the video is like a real video)\nHere is an output example:\nvisual quality: 3.2\ntemporal consistency: 2.7\ndynamic degree: 4.0\ntext-to-video alignment: 2.3\nfactual consistency: 1.8\n\nFor this video, the text prompt is \"A police car follows a hover sports car down the canyon freeway. Scifi. Cyberpunk.\",\nall the frames of video are as follows:\n\n<image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> " }, { "from": "gpt", "value": "visual quality: 1\n temporal consistency: 3\n dynamic degree: 2\n text-to-video alignment: 3\n factual consistency: 3\n" } ]

[ "0002689_00.jpg", "0002689_01.jpg", "0002689_02.jpg", "0002689_03.jpg", "0002689_04.jpg", "0002689_05.jpg", "0002689_06.jpg", "0002689_07.jpg", "0002689_08.jpg", "0002689_09.jpg", "0002689_10.jpg", "0002689_11.jpg", "0002689_12.jpg", "0002689_13.jpg", "0002689_14.jpg", "0002689_15.jpg", "0002689_16.jpg", "0002689_17.jpg", "0002689_18.jpg", "0002689_19.jpg", "0002689_20.jpg", "0002689_21.jpg", "0002689_22.jpg", "0002689_23.jpg" ]

[ { "from": "human", "value": "\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos,\nplease watch the following frames of a given video and see the text prompt for generating the video,\nthen give scores from 5 different dimensions:\n(1) visual quality: the quality of the video in terms of clearness, resolution, brightness, and color\n(2) temporal consistency, both the consistency of objects or humans and the smoothness of motion or movements\n(3) dynamic degree, the degree of dynamic changes\n(4) text-to-video alignment, the alignment between the text prompt and the video content\n(5) factual consistency, the consistency of the video content with the common-sense and factual knowledge\n\nfor each dimension, output a float number from 1.0 to 4.0,\nthe higher the number is, the better the video performs in that sub-score, \nthe lowest 1.0 means Bad, the highest 4.0 means Perfect/Real (the video is like a real video)\nHere is an output example:\nvisual quality: 3.2\ntemporal consistency: 2.7\ndynamic degree: 4.0\ntext-to-video alignment: 2.3\nfactual consistency: 1.8\n\nFor this video, the text prompt is \"The Guide imparts a mission to Maya, his piercing gaze reflecting the weight of the task. Maya, with a mix of determination and uncertainty, accepts the challenge, her backpack now containing newfound artifacts.\",\nall the frames of video are as follows:\n\n<image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> " }, { "from": "gpt", "value": "visual quality: 3\n temporal consistency: 3\n dynamic degree: 2\n text-to-video alignment: 3\n factual consistency: 3\n" } ]

[ "3000805_00.jpg", "3000805_01.jpg", "3000805_02.jpg", "3000805_03.jpg", "3000805_04.jpg", "3000805_05.jpg", "3000805_06.jpg", "3000805_07.jpg", "3000805_08.jpg", "3000805_09.jpg", "3000805_10.jpg", "3000805_11.jpg", "3000805_12.jpg", "3000805_13.jpg", "3000805_14.jpg", "3000805_15.jpg" ]

[ { "from": "human", "value": "\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos,\nplease watch the following frames of a given video and see the text prompt for generating the video,\nthen give scores from 5 different dimensions:\n(1) visual quality: the quality of the video in terms of clearness, resolution, brightness, and color\n(2) temporal consistency, both the consistency of objects or humans and the smoothness of motion or movements\n(3) dynamic degree, the degree of dynamic changes\n(4) text-to-video alignment, the alignment between the text prompt and the video content\n(5) factual consistency, the consistency of the video content with the common-sense and factual knowledge\n\nfor each dimension, output a float number from 1.0 to 4.0,\nthe higher the number is, the better the video performs in that sub-score, \nthe lowest 1.0 means Bad, the highest 4.0 means Perfect/Real (the video is like a real video)\nHere is an output example:\nvisual quality: 3.2\ntemporal consistency: 2.7\ndynamic degree: 4.0\ntext-to-video alignment: 2.3\nfactual consistency: 1.8\n\nFor this video, the text prompt is \"(Best quality, high: 1.2), pixar style, Disney character Bluey, (In front of a birthday cake), (the cake has two lit candles) pixar style\",\nall the frames of video are as follows:\n\n<image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> " }, { "from": "gpt", "value": "visual quality: 2\n temporal consistency: 3\n dynamic degree: 3\n text-to-video alignment: 3\n factual consistency: 3\n" } ]

[ "3002195_00.jpg", "3002195_01.jpg", "3002195_02.jpg", "3002195_03.jpg", "3002195_04.jpg", "3002195_05.jpg", "3002195_06.jpg", "3002195_07.jpg", "3002195_08.jpg", "3002195_09.jpg", "3002195_10.jpg", "3002195_11.jpg", "3002195_12.jpg", "3002195_13.jpg", "3002195_14.jpg", "3002195_15.jpg" ]

[ { "from": "human", "value": "\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos,\nplease watch the following frames of a given video and see the text prompt for generating the video,\nthen give scores from 5 different dimensions:\n(1) visual quality: the quality of the video in terms of clearness, resolution, brightness, and color\n(2) temporal consistency, both the consistency of objects or humans and the smoothness of motion or movements\n(3) dynamic degree, the degree of dynamic changes\n(4) text-to-video alignment, the alignment between the text prompt and the video content\n(5) factual consistency, the consistency of the video content with the common-sense and factual knowledge\n\nfor each dimension, output a float number from 1.0 to 4.0,\nthe higher the number is, the better the video performs in that sub-score, \nthe lowest 1.0 means Bad, the highest 4.0 means Perfect/Real (the video is like a real video)\nHere is an output example:\nvisual quality: 3.2\ntemporal consistency: 2.7\ndynamic degree: 4.0\ntext-to-video alignment: 2.3\nfactual consistency: 1.8\n\nFor this video, the text prompt is \"a black cat sleeping under a small car peacefully at night 9 pm\",\nall the frames of video are as follows:\n\n<image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> " }, { "from": "gpt", "value": "visual quality: 3\n temporal consistency: 3\n dynamic degree: 1\n text-to-video alignment: 3\n factual consistency: 2\n" } ]

[ "3001993_00.jpg", "3001993_01.jpg", "3001993_02.jpg", "3001993_03.jpg", "3001993_04.jpg", "3001993_05.jpg", "3001993_06.jpg", "3001993_07.jpg", "3001993_08.jpg", "3001993_09.jpg", "3001993_10.jpg", "3001993_11.jpg", "3001993_12.jpg", "3001993_13.jpg", "3001993_14.jpg", "3001993_15.jpg" ]

[ { "from": "human", "value": "\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos,\nplease watch the following frames of a given video and see the text prompt for generating the video,\nthen give scores from 5 different dimensions:\n(1) visual quality: the quality of the video in terms of clearness, resolution, brightness, and color\n(2) temporal consistency, both the consistency of objects or humans and the smoothness of motion or movements\n(3) dynamic degree, the degree of dynamic changes\n(4) text-to-video alignment, the alignment between the text prompt and the video content\n(5) factual consistency, the consistency of the video content with the common-sense and factual knowledge\n\nfor each dimension, output a float number from 1.0 to 4.0,\nthe higher the number is, the better the video performs in that sub-score, \nthe lowest 1.0 means Bad, the highest 4.0 means Perfect/Real (the video is like a real video)\nHere is an output example:\nvisual quality: 3.2\ntemporal consistency: 2.7\ndynamic degree: 4.0\ntext-to-video alignment: 2.3\nfactual consistency: 1.8\n\nFor this video, the text prompt is \"a father scolding his son for doing something wrong\",\nall the frames of video are as follows:\n\n<image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> " }, { "from": "gpt", "value": "visual quality: 2\n temporal consistency: 1\n dynamic degree: 3\n text-to-video alignment: 1\n factual consistency: 1\n" } ]

[ "3001699_00.jpg", "3001699_01.jpg", "3001699_02.jpg", "3001699_03.jpg", "3001699_04.jpg", "3001699_05.jpg", "3001699_06.jpg", "3001699_07.jpg", "3001699_08.jpg", "3001699_09.jpg", "3001699_10.jpg", "3001699_11.jpg", "3001699_12.jpg", "3001699_13.jpg", "3001699_14.jpg", "3001699_15.jpg" ]

[ { "from": "human", "value": "\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos,\nplease watch the following frames of a given video and see the text prompt for generating the video,\nthen give scores from 5 different dimensions:\n(1) visual quality: the quality of the video in terms of clearness, resolution, brightness, and color\n(2) temporal consistency, both the consistency of objects or humans and the smoothness of motion or movements\n(3) dynamic degree, the degree of dynamic changes\n(4) text-to-video alignment, the alignment between the text prompt and the video content\n(5) factual consistency, the consistency of the video content with the common-sense and factual knowledge\n\nfor each dimension, output a float number from 1.0 to 4.0,\nthe higher the number is, the better the video performs in that sub-score, \nthe lowest 1.0 means Bad, the highest 4.0 means Perfect/Real (the video is like a real video)\nHere is an output example:\nvisual quality: 3.2\ntemporal consistency: 2.7\ndynamic degree: 4.0\ntext-to-video alignment: 2.3\nfactual consistency: 1.8\n\nFor this video, the text prompt is \"epic shots of the computer game tournament venue, filled with cheering crowds and flashing lights\",\nall the frames of video are as follows:\n\n<image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> " }, { "from": "gpt", "value": "visual quality: 2\n temporal consistency: 1\n dynamic degree: 3\n text-to-video alignment: 1\n factual consistency: 1\n" } ]

[ "2000570_00.jpg", "2000570_01.jpg", "2000570_02.jpg", "2000570_03.jpg", "2000570_04.jpg", "2000570_05.jpg", "2000570_06.jpg", "2000570_07.jpg", "2000570_08.jpg", "2000570_09.jpg", "2000570_10.jpg", "2000570_11.jpg", "2000570_12.jpg", "2000570_13.jpg", "2000570_14.jpg", "2000570_15.jpg" ]

[ { "from": "human", "value": "\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos,\nplease watch the following frames of a given video and see the text prompt for generating the video,\nthen give scores from 5 different dimensions:\n(1) visual quality: the quality of the video in terms of clearness, resolution, brightness, and color\n(2) temporal consistency, both the consistency of objects or humans and the smoothness of motion or movements\n(3) dynamic degree, the degree of dynamic changes\n(4) text-to-video alignment, the alignment between the text prompt and the video content\n(5) factual consistency, the consistency of the video content with the common-sense and factual knowledge\n\nfor each dimension, output a float number from 1.0 to 4.0,\nthe higher the number is, the better the video performs in that sub-score, \nthe lowest 1.0 means Bad, the highest 4.0 means Perfect/Real (the video is like a real video)\nHere is an output example:\nvisual quality: 3.2\ntemporal consistency: 2.7\ndynamic degree: 4.0\ntext-to-video alignment: 2.3\nfactual consistency: 1.8\n\nFor this video, the text prompt is \"A red heart shape appearing out of a foggy background\",\nall the frames of video are as follows:\n\n<image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> " }, { "from": "gpt", "value": "visual quality: 3\n temporal consistency: 2\n dynamic degree: 3\n text-to-video alignment: 3\n factual consistency: 3\n" } ]

[ "3002949_00.jpg", "3002949_01.jpg", "3002949_02.jpg", "3002949_03.jpg", "3002949_04.jpg", "3002949_05.jpg", "3002949_06.jpg", "3002949_07.jpg", "3002949_08.jpg", "3002949_09.jpg", "3002949_10.jpg", "3002949_11.jpg", "3002949_12.jpg", "3002949_13.jpg", "3002949_14.jpg", "3002949_15.jpg" ]

[ { "from": "human", "value": "\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos,\nplease watch the following frames of a given video and see the text prompt for generating the video,\nthen give scores from 5 different dimensions:\n(1) visual quality: the quality of the video in terms of clearness, resolution, brightness, and color\n(2) temporal consistency, both the consistency of objects or humans and the smoothness of motion or movements\n(3) dynamic degree, the degree of dynamic changes\n(4) text-to-video alignment, the alignment between the text prompt and the video content\n(5) factual consistency, the consistency of the video content with the common-sense and factual knowledge\n\nfor each dimension, output a float number from 1.0 to 4.0,\nthe higher the number is, the better the video performs in that sub-score, \nthe lowest 1.0 means Bad, the highest 4.0 means Perfect/Real (the video is like a real video)\nHere is an output example:\nvisual quality: 3.2\ntemporal consistency: 2.7\ndynamic degree: 4.0\ntext-to-video alignment: 2.3\nfactual consistency: 1.8\n\nFor this video, the text prompt is \"A karthus animation for league of legends\",\nall the frames of video are as follows:\n\n<image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> " }, { "from": "gpt", "value": "visual quality: 3\n temporal consistency: 2\n dynamic degree: 3\n text-to-video alignment: 2\n factual consistency: 2\n" } ]

[ "2002929_00.jpg", "2002929_01.jpg", "2002929_02.jpg", "2002929_03.jpg", "2002929_04.jpg", "2002929_05.jpg", "2002929_06.jpg", "2002929_07.jpg", "2002929_08.jpg", "2002929_09.jpg", "2002929_10.jpg", "2002929_11.jpg", "2002929_12.jpg", "2002929_13.jpg", "2002929_14.jpg", "2002929_15.jpg" ]

[ { "from": "human", "value": "\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos,\nplease watch the following frames of a given video and see the text prompt for generating the video,\nthen give scores from 5 different dimensions:\n(1) visual quality: the quality of the video in terms of clearness, resolution, brightness, and color\n(2) temporal consistency, both the consistency of objects or humans and the smoothness of motion or movements\n(3) dynamic degree, the degree of dynamic changes\n(4) text-to-video alignment, the alignment between the text prompt and the video content\n(5) factual consistency, the consistency of the video content with the common-sense and factual knowledge\n\nfor each dimension, output a float number from 1.0 to 4.0,\nthe higher the number is, the better the video performs in that sub-score, \nthe lowest 1.0 means Bad, the highest 4.0 means Perfect/Real (the video is like a real video)\nHere is an output example:\nvisual quality: 3.2\ntemporal consistency: 2.7\ndynamic degree: 4.0\ntext-to-video alignment: 2.3\nfactual consistency: 1.8\n\nFor this video, the text prompt is \"a loop video of dancing breakdance one dancer contest on white background and zoom out and rotate camaera point of view\",\nall the frames of video are as follows:\n\n<image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> " }, { "from": "gpt", "value": "visual quality: 3\n temporal consistency: 2\n dynamic degree: 3\n text-to-video alignment: 3\n factual consistency: 3\n" } ]

[ "6001928_00.jpg", "6001928_01.jpg", "6001928_02.jpg", "6001928_03.jpg", "6001928_04.jpg", "6001928_05.jpg", "6001928_06.jpg", "6001928_07.jpg", "6001928_08.jpg", "6001928_09.jpg", "6001928_10.jpg", "6001928_11.jpg", "6001928_12.jpg", "6001928_13.jpg", "6001928_14.jpg", "6001928_15.jpg" ]

[ { "from": "human", "value": "\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos,\nplease watch the following frames of a given video and see the text prompt for generating the video,\nthen give scores from 5 different dimensions:\n(1) visual quality: the quality of the video in terms of clearness, resolution, brightness, and color\n(2) temporal consistency, both the consistency of objects or humans and the smoothness of motion or movements\n(3) dynamic degree, the degree of dynamic changes\n(4) text-to-video alignment, the alignment between the text prompt and the video content\n(5) factual consistency, the consistency of the video content with the common-sense and factual knowledge\n\nfor each dimension, output a float number from 1.0 to 4.0,\nthe higher the number is, the better the video performs in that sub-score, \nthe lowest 1.0 means Bad, the highest 4.0 means Perfect/Real (the video is like a real video)\nHere is an output example:\nvisual quality: 3.2\ntemporal consistency: 2.7\ndynamic degree: 4.0\ntext-to-video alignment: 2.3\nfactual consistency: 1.8\n\nFor this video, the text prompt is \"random chrome objects floating in a wavy moving dark red goo\",\nall the frames of video are as follows:\n\n<image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> " }, { "from": "gpt", "value": "visual quality: 1\n temporal consistency: 2\n dynamic degree: 3\n text-to-video alignment: 3\n factual consistency: 1\n" } ]

[ "7001065_00.jpg", "7001065_01.jpg", "7001065_02.jpg", "7001065_03.jpg", "7001065_04.jpg", "7001065_05.jpg", "7001065_06.jpg", "7001065_07.jpg" ]

[ { "from": "human", "value": "\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos,\nplease watch the following frames of a given video and see the text prompt for generating the video,\nthen give scores from 5 different dimensions:\n(1) visual quality: the quality of the video in terms of clearness, resolution, brightness, and color\n(2) temporal consistency, both the consistency of objects or humans and the smoothness of motion or movements\n(3) dynamic degree, the degree of dynamic changes\n(4) text-to-video alignment, the alignment between the text prompt and the video content\n(5) factual consistency, the consistency of the video content with the common-sense and factual knowledge\n\nfor each dimension, output a float number from 1.0 to 4.0,\nthe higher the number is, the better the video performs in that sub-score, \nthe lowest 1.0 means Bad, the highest 4.0 means Perfect/Real (the video is like a real video)\nHere is an output example:\nvisual quality: 3.2\ntemporal consistency: 2.7\ndynamic degree: 4.0\ntext-to-video alignment: 2.3\nfactual consistency: 1.8\n\nFor this video, the text prompt is \"a horse running in a long beach\",\nall the frames of video are as follows:\n\n<image> <image> <image> <image> <image> <image> <image> <image> " }, { "from": "gpt", "value": "visual quality: 2\n temporal consistency: 2\n dynamic degree: 3\n text-to-video alignment: 2\n factual consistency: 2\n" } ]

[ "3001912_00.jpg", "3001912_01.jpg", "3001912_02.jpg", "3001912_03.jpg", "3001912_04.jpg", "3001912_05.jpg", "3001912_06.jpg", "3001912_07.jpg", "3001912_08.jpg", "3001912_09.jpg", "3001912_10.jpg", "3001912_11.jpg", "3001912_12.jpg", "3001912_13.jpg", "3001912_14.jpg", "3001912_15.jpg" ]

[ { "from": "human", "value": "\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos,\nplease watch the following frames of a given video and see the text prompt for generating the video,\nthen give scores from 5 different dimensions:\n(1) visual quality: the quality of the video in terms of clearness, resolution, brightness, and color\n(2) temporal consistency, both the consistency of objects or humans and the smoothness of motion or movements\n(3) dynamic degree, the degree of dynamic changes\n(4) text-to-video alignment, the alignment between the text prompt and the video content\n(5) factual consistency, the consistency of the video content with the common-sense and factual knowledge\n\nfor each dimension, output a float number from 1.0 to 4.0,\nthe higher the number is, the better the video performs in that sub-score, \nthe lowest 1.0 means Bad, the highest 4.0 means Perfect/Real (the video is like a real video)\nHere is an output example:\nvisual quality: 3.2\ntemporal consistency: 2.7\ndynamic degree: 4.0\ntext-to-video alignment: 2.3\nfactual consistency: 1.8\n\nFor this video, the text prompt is \"Massive computers and habitable structures can now be seen on the inner surface of the sphere, which has a radius larger than the orbit of Mars.\",\nall the frames of video are as follows:\n\n<image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> " }, { "from": "gpt", "value": "visual quality: 2\n temporal consistency: 3\n dynamic degree: 3\n text-to-video alignment: 3\n factual consistency: 3\n" } ]

[ "4001922_00.jpg", "4001922_01.jpg", "4001922_02.jpg", "4001922_03.jpg", "4001922_04.jpg", "4001922_05.jpg", "4001922_06.jpg", "4001922_07.jpg", "4001922_08.jpg", "4001922_09.jpg", "4001922_10.jpg", "4001922_11.jpg", "4001922_12.jpg", "4001922_13.jpg", "4001922_14.jpg", "4001922_15.jpg" ]

[ { "from": "human", "value": "\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos,\nplease watch the following frames of a given video and see the text prompt for generating the video,\nthen give scores from 5 different dimensions:\n(1) visual quality: the quality of the video in terms of clearness, resolution, brightness, and color\n(2) temporal consistency, both the consistency of objects or humans and the smoothness of motion or movements\n(3) dynamic degree, the degree of dynamic changes\n(4) text-to-video alignment, the alignment between the text prompt and the video content\n(5) factual consistency, the consistency of the video content with the common-sense and factual knowledge\n\nfor each dimension, output a float number from 1.0 to 4.0,\nthe higher the number is, the better the video performs in that sub-score, \nthe lowest 1.0 means Bad, the highest 4.0 means Perfect/Real (the video is like a real video)\nHere is an output example:\nvisual quality: 3.2\ntemporal consistency: 2.7\ndynamic degree: 4.0\ntext-to-video alignment: 2.3\nfactual consistency: 1.8\n\nFor this video, the text prompt is \"A robot running on a treadmill\",\nall the frames of video are as follows:\n\n<image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> " }, { "from": "gpt", "value": "visual quality: 3\n temporal consistency: 2\n dynamic degree: 3\n text-to-video alignment: 3\n factual consistency: 3\n" } ]

[ "1001333_00.jpg", "1001333_01.jpg", "1001333_02.jpg", "1001333_03.jpg", "1001333_04.jpg", "1001333_05.jpg", "1001333_06.jpg", "1001333_07.jpg", "1001333_08.jpg", "1001333_09.jpg", "1001333_10.jpg", "1001333_11.jpg", "1001333_12.jpg", "1001333_13.jpg", "1001333_14.jpg", "1001333_15.jpg" ]

[ { "from": "human", "value": "\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos,\nplease watch the following frames of a given video and see the text prompt for generating the video,\nthen give scores from 5 different dimensions:\n(1) visual quality: the quality of the video in terms of clearness, resolution, brightness, and color\n(2) temporal consistency, both the consistency of objects or humans and the smoothness of motion or movements\n(3) dynamic degree, the degree of dynamic changes\n(4) text-to-video alignment, the alignment between the text prompt and the video content\n(5) factual consistency, the consistency of the video content with the common-sense and factual knowledge\n\nfor each dimension, output a float number from 1.0 to 4.0,\nthe higher the number is, the better the video performs in that sub-score, \nthe lowest 1.0 means Bad, the highest 4.0 means Perfect/Real (the video is like a real video)\nHere is an output example:\nvisual quality: 3.2\ntemporal consistency: 2.7\ndynamic degree: 4.0\ntext-to-video alignment: 2.3\nfactual consistency: 1.8\n\nFor this video, the text prompt is \"a big red dog drinks coffee in the metropolis\",\nall the frames of video are as follows:\n\n<image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> " }, { "from": "gpt", "value": "visual quality: 1\n temporal consistency: 1\n dynamic degree: 3\n text-to-video alignment: 2\n factual consistency: 1\n" } ]

[ "3002323_00.jpg", "3002323_01.jpg", "3002323_02.jpg", "3002323_03.jpg", "3002323_04.jpg", "3002323_05.jpg", "3002323_06.jpg", "3002323_07.jpg", "3002323_08.jpg", "3002323_09.jpg", "3002323_10.jpg", "3002323_11.jpg", "3002323_12.jpg", "3002323_13.jpg", "3002323_14.jpg", "3002323_15.jpg" ]

[ { "from": "human", "value": "\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos,\nplease watch the following frames of a given video and see the text prompt for generating the video,\nthen give scores from 5 different dimensions:\n(1) visual quality: the quality of the video in terms of clearness, resolution, brightness, and color\n(2) temporal consistency, both the consistency of objects or humans and the smoothness of motion or movements\n(3) dynamic degree, the degree of dynamic changes\n(4) text-to-video alignment, the alignment between the text prompt and the video content\n(5) factual consistency, the consistency of the video content with the common-sense and factual knowledge\n\nfor each dimension, output a float number from 1.0 to 4.0,\nthe higher the number is, the better the video performs in that sub-score, \nthe lowest 1.0 means Bad, the highest 4.0 means Perfect/Real (the video is like a real video)\nHere is an output example:\nvisual quality: 3.2\ntemporal consistency: 2.7\ndynamic degree: 4.0\ntext-to-video alignment: 2.3\nfactual consistency: 1.8\n\nFor this video, the text prompt is \"smart technology, wind turbines, clean energy, sustainable minimilastic , image,\",\nall the frames of video are as follows:\n\n<image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> " }, { "from": "gpt", "value": "visual quality: 2\n temporal consistency: 1\n dynamic degree: 3\n text-to-video alignment: 2\n factual consistency: 2\n" } ]

[ "3001947_00.jpg", "3001947_01.jpg", "3001947_02.jpg", "3001947_03.jpg", "3001947_04.jpg", "3001947_05.jpg", "3001947_06.jpg", "3001947_07.jpg", "3001947_08.jpg", "3001947_09.jpg", "3001947_10.jpg", "3001947_11.jpg", "3001947_12.jpg", "3001947_13.jpg", "3001947_14.jpg", "3001947_15.jpg" ]

[ { "from": "human", "value": "\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos,\nplease watch the following frames of a given video and see the text prompt for generating the video,\nthen give scores from 5 different dimensions:\n(1) visual quality: the quality of the video in terms of clearness, resolution, brightness, and color\n(2) temporal consistency, both the consistency of objects or humans and the smoothness of motion or movements\n(3) dynamic degree, the degree of dynamic changes\n(4) text-to-video alignment, the alignment between the text prompt and the video content\n(5) factual consistency, the consistency of the video content with the common-sense and factual knowledge\n\nfor each dimension, output a float number from 1.0 to 4.0,\nthe higher the number is, the better the video performs in that sub-score, \nthe lowest 1.0 means Bad, the highest 4.0 means Perfect/Real (the video is like a real video)\nHere is an output example:\nvisual quality: 3.2\ntemporal consistency: 2.7\ndynamic degree: 4.0\ntext-to-video alignment: 2.3\nfactual consistency: 1.8\n\nFor this video, the text prompt is \"Holographic MachineEntity femine portrait techniclor projection of a young woman\",\nall the frames of video are as follows:\n\n<image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> " }, { "from": "gpt", "value": "visual quality: 2\n temporal consistency: 1\n dynamic degree: 3\n text-to-video alignment: 1\n factual consistency: 1\n" } ]

[ "3000016_00.jpg", "3000016_01.jpg", "3000016_02.jpg", "3000016_03.jpg", "3000016_04.jpg", "3000016_05.jpg", "3000016_06.jpg", "3000016_07.jpg", "3000016_08.jpg", "3000016_09.jpg", "3000016_10.jpg", "3000016_11.jpg", "3000016_12.jpg", "3000016_13.jpg", "3000016_14.jpg", "3000016_15.jpg" ]

[ { "from": "human", "value": "\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos,\nplease watch the following frames of a given video and see the text prompt for generating the video,\nthen give scores from 5 different dimensions:\n(1) visual quality: the quality of the video in terms of clearness, resolution, brightness, and color\n(2) temporal consistency, both the consistency of objects or humans and the smoothness of motion or movements\n(3) dynamic degree, the degree of dynamic changes\n(4) text-to-video alignment, the alignment between the text prompt and the video content\n(5) factual consistency, the consistency of the video content with the common-sense and factual knowledge\n\nfor each dimension, output a float number from 1.0 to 4.0,\nthe higher the number is, the better the video performs in that sub-score, \nthe lowest 1.0 means Bad, the highest 4.0 means Perfect/Real (the video is like a real video)\nHere is an output example:\nvisual quality: 3.2\ntemporal consistency: 2.7\ndynamic degree: 4.0\ntext-to-video alignment: 2.3\nfactual consistency: 1.8\n\nFor this video, the text prompt is \"silent land with mountains and rocks and hills\",\nall the frames of video are as follows:\n\n<image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> " }, { "from": "gpt", "value": "visual quality: 2\n temporal consistency: 3\n dynamic degree: 3\n text-to-video alignment: 3\n factual consistency: 3\n" } ]

[ "0002607_00.jpg", "0002607_01.jpg", "0002607_02.jpg", "0002607_03.jpg", "0002607_04.jpg", "0002607_05.jpg", "0002607_06.jpg", "0002607_07.jpg", "0002607_08.jpg", "0002607_09.jpg", "0002607_10.jpg", "0002607_11.jpg", "0002607_12.jpg", "0002607_13.jpg", "0002607_14.jpg", "0002607_15.jpg", "0002607_16.jpg", "0002607_17.jpg", "0002607_18.jpg", "0002607_19.jpg", "0002607_20.jpg", "0002607_21.jpg", "0002607_22.jpg", "0002607_23.jpg" ]

[ { "from": "human", "value": "\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos,\nplease watch the following frames of a given video and see the text prompt for generating the video,\nthen give scores from 5 different dimensions:\n(1) visual quality: the quality of the video in terms of clearness, resolution, brightness, and color\n(2) temporal consistency, both the consistency of objects or humans and the smoothness of motion or movements\n(3) dynamic degree, the degree of dynamic changes\n(4) text-to-video alignment, the alignment between the text prompt and the video content\n(5) factual consistency, the consistency of the video content with the common-sense and factual knowledge\n\nfor each dimension, output a float number from 1.0 to 4.0,\nthe higher the number is, the better the video performs in that sub-score, \nthe lowest 1.0 means Bad, the highest 4.0 means Perfect/Real (the video is like a real video)\nHere is an output example:\nvisual quality: 3.2\ntemporal consistency: 2.7\ndynamic degree: 4.0\ntext-to-video alignment: 2.3\nfactual consistency: 1.8\n\nFor this video, the text prompt is \"students in the photography studio shooting a magazine brief\",\nall the frames of video are as follows:\n\n<image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> " }, { "from": "gpt", "value": "visual quality: 3\n temporal consistency: 3\n dynamic degree: 2\n text-to-video alignment: 3\n factual consistency: 3\n" } ]

[ "2002044_00.jpg", "2002044_01.jpg", "2002044_02.jpg", "2002044_03.jpg", "2002044_04.jpg", "2002044_05.jpg", "2002044_06.jpg", "2002044_07.jpg", "2002044_08.jpg", "2002044_09.jpg", "2002044_10.jpg", "2002044_11.jpg", "2002044_12.jpg", "2002044_13.jpg", "2002044_14.jpg", "2002044_15.jpg" ]

[ { "from": "human", "value": "\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos,\nplease watch the following frames of a given video and see the text prompt for generating the video,\nthen give scores from 5 different dimensions:\n(1) visual quality: the quality of the video in terms of clearness, resolution, brightness, and color\n(2) temporal consistency, both the consistency of objects or humans and the smoothness of motion or movements\n(3) dynamic degree, the degree of dynamic changes\n(4) text-to-video alignment, the alignment between the text prompt and the video content\n(5) factual consistency, the consistency of the video content with the common-sense and factual knowledge\n\nfor each dimension, output a float number from 1.0 to 4.0,\nthe higher the number is, the better the video performs in that sub-score, \nthe lowest 1.0 means Bad, the highest 4.0 means Perfect/Real (the video is like a real video)\nHere is an output example:\nvisual quality: 3.2\ntemporal consistency: 2.7\ndynamic degree: 4.0\ntext-to-video alignment: 2.3\nfactual consistency: 1.8\n\nFor this video, the text prompt is \"16:9 resolution 2048, in comic book style, Raul Seixas sleeps in your bed in a dark room full of electronic devices, such as TV, cell phone, computer, tablet, etc.\",\nall the frames of video are as follows:\n\n<image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> " }, { "from": "gpt", "value": "visual quality: 3\n temporal consistency: 3\n dynamic degree: 2\n text-to-video alignment: 2\n factual consistency: 3\n" } ]

[ "8000615_00.jpg", "8000615_01.jpg", "8000615_02.jpg", "8000615_03.jpg", "8000615_04.jpg", "8000615_05.jpg", "8000615_06.jpg", "8000615_07.jpg", "8000615_08.jpg", "8000615_09.jpg", "8000615_10.jpg", "8000615_11.jpg", "8000615_12.jpg", "8000615_13.jpg", "8000615_14.jpg", "8000615_15.jpg", "8000615_16.jpg", "8000615_17.jpg", "8000615_18.jpg", "8000615_19.jpg", "8000615_20.jpg", "8000615_21.jpg", "8000615_22.jpg", "8000615_23.jpg" ]

[ { "from": "human", "value": "\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos,\nplease watch the following frames of a given video and see the text prompt for generating the video,\nthen give scores from 5 different dimensions:\n(1) visual quality: the quality of the video in terms of clearness, resolution, brightness, and color\n(2) temporal consistency, both the consistency of objects or humans and the smoothness of motion or movements\n(3) dynamic degree, the degree of dynamic changes\n(4) text-to-video alignment, the alignment between the text prompt and the video content\n(5) factual consistency, the consistency of the video content with the common-sense and factual knowledge\n\nfor each dimension, output a float number from 1.0 to 4.0,\nthe higher the number is, the better the video performs in that sub-score, \nthe lowest 1.0 means Bad, the highest 4.0 means Perfect/Real (the video is like a real video)\nHere is an output example:\nvisual quality: 3.2\ntemporal consistency: 2.7\ndynamic degree: 4.0\ntext-to-video alignment: 2.3\nfactual consistency: 1.8\n\nFor this video, the text prompt is \"The wrght brothers building the first plane\",\nall the frames of video are as follows:\n\n<image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> " }, { "from": "gpt", "value": "visual quality: 1\n temporal consistency: 1\n dynamic degree: 3\n text-to-video alignment: 2\n factual consistency: 1\n" } ]

[ "5000428_00.jpg", "5000428_01.jpg", "5000428_02.jpg", "5000428_03.jpg", "5000428_04.jpg", "5000428_05.jpg", "5000428_06.jpg", "5000428_07.jpg", "5000428_08.jpg", "5000428_09.jpg", "5000428_10.jpg", "5000428_11.jpg", "5000428_12.jpg", "5000428_13.jpg", "5000428_14.jpg", "5000428_15.jpg" ]

[ { "from": "human", "value": "\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos,\nplease watch the following frames of a given video and see the text prompt for generating the video,\nthen give scores from 5 different dimensions:\n(1) visual quality: the quality of the video in terms of clearness, resolution, brightness, and color\n(2) temporal consistency, both the consistency of objects or humans and the smoothness of motion or movements\n(3) dynamic degree, the degree of dynamic changes\n(4) text-to-video alignment, the alignment between the text prompt and the video content\n(5) factual consistency, the consistency of the video content with the common-sense and factual knowledge\n\nfor each dimension, output a float number from 1.0 to 4.0,\nthe higher the number is, the better the video performs in that sub-score, \nthe lowest 1.0 means Bad, the highest 4.0 means Perfect/Real (the video is like a real video)\nHere is an output example:\nvisual quality: 3.2\ntemporal consistency: 2.7\ndynamic degree: 4.0\ntext-to-video alignment: 2.3\nfactual consistency: 1.8\n\nFor this video, the text prompt is \"velvet vixen reads the playboy magazin in LV sunglasses\",\nall the frames of video are as follows:\n\n<image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> " }, { "from": "gpt", "value": "visual quality: 2\n temporal consistency: 1\n dynamic degree: 3\n text-to-video alignment: 3\n factual consistency: 1\n" } ]

[ "0002270_00.jpg", "0002270_01.jpg", "0002270_02.jpg", "0002270_03.jpg", "0002270_04.jpg", "0002270_05.jpg", "0002270_06.jpg", "0002270_07.jpg", "0002270_08.jpg", "0002270_09.jpg", "0002270_10.jpg", "0002270_11.jpg", "0002270_12.jpg", "0002270_13.jpg", "0002270_14.jpg", "0002270_15.jpg", "0002270_16.jpg", "0002270_17.jpg", "0002270_18.jpg", "0002270_19.jpg", "0002270_20.jpg", "0002270_21.jpg", "0002270_22.jpg", "0002270_23.jpg" ]

[ { "from": "human", "value": "\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos,\nplease watch the following frames of a given video and see the text prompt for generating the video,\nthen give scores from 5 different dimensions:\n(1) visual quality: the quality of the video in terms of clearness, resolution, brightness, and color\n(2) temporal consistency, both the consistency of objects or humans and the smoothness of motion or movements\n(3) dynamic degree, the degree of dynamic changes\n(4) text-to-video alignment, the alignment between the text prompt and the video content\n(5) factual consistency, the consistency of the video content with the common-sense and factual knowledge\n\nfor each dimension, output a float number from 1.0 to 4.0,\nthe higher the number is, the better the video performs in that sub-score, \nthe lowest 1.0 means Bad, the highest 4.0 means Perfect/Real (the video is like a real video)\nHere is an output example:\nvisual quality: 3.2\ntemporal consistency: 2.7\ndynamic degree: 4.0\ntext-to-video alignment: 2.3\nfactual consistency: 1.8\n\nFor this video, the text prompt is \"tesla model is moving fast in a futuristic crystal city\",\nall the frames of video are as follows:\n\n<image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> " }, { "from": "gpt", "value": "visual quality: 3\n temporal consistency: 3\n dynamic degree: 2\n text-to-video alignment: 3\n factual consistency: 3\n" } ]

[ "3002930_00.jpg", "3002930_01.jpg", "3002930_02.jpg", "3002930_03.jpg", "3002930_04.jpg", "3002930_05.jpg", "3002930_06.jpg", "3002930_07.jpg", "3002930_08.jpg", "3002930_09.jpg", "3002930_10.jpg", "3002930_11.jpg", "3002930_12.jpg", "3002930_13.jpg", "3002930_14.jpg", "3002930_15.jpg" ]

[ { "from": "human", "value": "\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos,\nplease watch the following frames of a given video and see the text prompt for generating the video,\nthen give scores from 5 different dimensions:\n(1) visual quality: the quality of the video in terms of clearness, resolution, brightness, and color\n(2) temporal consistency, both the consistency of objects or humans and the smoothness of motion or movements\n(3) dynamic degree, the degree of dynamic changes\n(4) text-to-video alignment, the alignment between the text prompt and the video content\n(5) factual consistency, the consistency of the video content with the common-sense and factual knowledge\n\nfor each dimension, output a float number from 1.0 to 4.0,\nthe higher the number is, the better the video performs in that sub-score, \nthe lowest 1.0 means Bad, the highest 4.0 means Perfect/Real (the video is like a real video)\nHere is an output example:\nvisual quality: 3.2\ntemporal consistency: 2.7\ndynamic degree: 4.0\ntext-to-video alignment: 2.3\nfactual consistency: 1.8\n\nFor this video, the text prompt is \"happy thick woman in the beach running\",\nall the frames of video are as follows:\n\n<image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> " }, { "from": "gpt", "value": "visual quality: 2\n temporal consistency: 1\n dynamic degree: 3\n text-to-video alignment: 1\n factual consistency: 1\n" } ]

[ "0002892_00.jpg", "0002892_01.jpg", "0002892_02.jpg", "0002892_03.jpg", "0002892_04.jpg", "0002892_05.jpg", "0002892_06.jpg", "0002892_07.jpg", "0002892_08.jpg", "0002892_09.jpg", "0002892_10.jpg", "0002892_11.jpg", "0002892_12.jpg", "0002892_13.jpg", "0002892_14.jpg", "0002892_15.jpg", "0002892_16.jpg", "0002892_17.jpg", "0002892_18.jpg", "0002892_19.jpg", "0002892_20.jpg", "0002892_21.jpg", "0002892_22.jpg", "0002892_23.jpg" ]

[ { "from": "human", "value": "\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos,\nplease watch the following frames of a given video and see the text prompt for generating the video,\nthen give scores from 5 different dimensions:\n(1) visual quality: the quality of the video in terms of clearness, resolution, brightness, and color\n(2) temporal consistency, both the consistency of objects or humans and the smoothness of motion or movements\n(3) dynamic degree, the degree of dynamic changes\n(4) text-to-video alignment, the alignment between the text prompt and the video content\n(5) factual consistency, the consistency of the video content with the common-sense and factual knowledge\n\nfor each dimension, output a float number from 1.0 to 4.0,\nthe higher the number is, the better the video performs in that sub-score, \nthe lowest 1.0 means Bad, the highest 4.0 means Perfect/Real (the video is like a real video)\nHere is an output example:\nvisual quality: 3.2\ntemporal consistency: 2.7\ndynamic degree: 4.0\ntext-to-video alignment: 2.3\nfactual consistency: 1.8\n\nFor this video, the text prompt is \"Majestic Asian Elephants moving gracefully through vast landscapes.\",\nall the frames of video are as follows:\n\n<image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> " }, { "from": "gpt", "value": "visual quality: 3\n temporal consistency: 3\n dynamic degree: 2\n text-to-video alignment: 3\n factual consistency: 3\n" } ]

[ "1000255_00.jpg", "1000255_01.jpg", "1000255_02.jpg", "1000255_03.jpg", "1000255_04.jpg", "1000255_05.jpg", "1000255_06.jpg", "1000255_07.jpg", "1000255_08.jpg", "1000255_09.jpg", "1000255_10.jpg", "1000255_11.jpg", "1000255_12.jpg", "1000255_13.jpg", "1000255_14.jpg", "1000255_15.jpg" ]

[ { "from": "human", "value": "\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos,\nplease watch the following frames of a given video and see the text prompt for generating the video,\nthen give scores from 5 different dimensions:\n(1) visual quality: the quality of the video in terms of clearness, resolution, brightness, and color\n(2) temporal consistency, both the consistency of objects or humans and the smoothness of motion or movements\n(3) dynamic degree, the degree of dynamic changes\n(4) text-to-video alignment, the alignment between the text prompt and the video content\n(5) factual consistency, the consistency of the video content with the common-sense and factual knowledge\n\nfor each dimension, output a float number from 1.0 to 4.0,\nthe higher the number is, the better the video performs in that sub-score, \nthe lowest 1.0 means Bad, the highest 4.0 means Perfect/Real (the video is like a real video)\nHere is an output example:\nvisual quality: 3.2\ntemporal consistency: 2.7\ndynamic degree: 4.0\ntext-to-video alignment: 2.3\nfactual consistency: 1.8\n\nFor this video, the text prompt is \"Each thread is spun with mathematical precision, forming a remarkably strong structure.\",\nall the frames of video are as follows:\n\n<image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> " }, { "from": "gpt", "value": "visual quality: 1\n temporal consistency: 2\n dynamic degree: 2\n text-to-video alignment: 2\n factual consistency: 1\n" } ]

[ "0002606_00.jpg", "0002606_01.jpg", "0002606_02.jpg", "0002606_03.jpg", "0002606_04.jpg", "0002606_05.jpg", "0002606_06.jpg", "0002606_07.jpg", "0002606_08.jpg", "0002606_09.jpg", "0002606_10.jpg", "0002606_11.jpg", "0002606_12.jpg", "0002606_13.jpg", "0002606_14.jpg", "0002606_15.jpg", "0002606_16.jpg", "0002606_17.jpg", "0002606_18.jpg", "0002606_19.jpg", "0002606_20.jpg", "0002606_21.jpg", "0002606_22.jpg", "0002606_23.jpg" ]

[ { "from": "human", "value": "\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos,\nplease watch the following frames of a given video and see the text prompt for generating the video,\nthen give scores from 5 different dimensions:\n(1) visual quality: the quality of the video in terms of clearness, resolution, brightness, and color\n(2) temporal consistency, both the consistency of objects or humans and the smoothness of motion or movements\n(3) dynamic degree, the degree of dynamic changes\n(4) text-to-video alignment, the alignment between the text prompt and the video content\n(5) factual consistency, the consistency of the video content with the common-sense and factual knowledge\n\nfor each dimension, output a float number from 1.0 to 4.0,\nthe higher the number is, the better the video performs in that sub-score, \nthe lowest 1.0 means Bad, the highest 4.0 means Perfect/Real (the video is like a real video)\nHere is an output example:\nvisual quality: 3.2\ntemporal consistency: 2.7\ndynamic degree: 4.0\ntext-to-video alignment: 2.3\nfactual consistency: 1.8\n\nFor this video, the text prompt is \"a short clip showing mountains of Nepal, with animals like polar bear\",\nall the frames of video are as follows:\n\n<image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> " }, { "from": "gpt", "value": "visual quality: 3\n temporal consistency: 3\n dynamic degree: 2\n text-to-video alignment: 2\n factual consistency: 3\n" } ]

[ "0002542_00.jpg", "0002542_01.jpg", "0002542_02.jpg", "0002542_03.jpg", "0002542_04.jpg", "0002542_05.jpg", "0002542_06.jpg", "0002542_07.jpg", "0002542_08.jpg", "0002542_09.jpg", "0002542_10.jpg", "0002542_11.jpg", "0002542_12.jpg", "0002542_13.jpg", "0002542_14.jpg", "0002542_15.jpg", "0002542_16.jpg", "0002542_17.jpg", "0002542_18.jpg", "0002542_19.jpg", "0002542_20.jpg", "0002542_21.jpg", "0002542_22.jpg", "0002542_23.jpg" ]

[ { "from": "human", "value": "\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos,\nplease watch the following frames of a given video and see the text prompt for generating the video,\nthen give scores from 5 different dimensions:\n(1) visual quality: the quality of the video in terms of clearness, resolution, brightness, and color\n(2) temporal consistency, both the consistency of objects or humans and the smoothness of motion or movements\n(3) dynamic degree, the degree of dynamic changes\n(4) text-to-video alignment, the alignment between the text prompt and the video content\n(5) factual consistency, the consistency of the video content with the common-sense and factual knowledge\n\nfor each dimension, output a float number from 1.0 to 4.0,\nthe higher the number is, the better the video performs in that sub-score, \nthe lowest 1.0 means Bad, the highest 4.0 means Perfect/Real (the video is like a real video)\nHere is an output example:\nvisual quality: 3.2\ntemporal consistency: 2.7\ndynamic degree: 4.0\ntext-to-video alignment: 2.3\nfactual consistency: 1.8\n\nFor this video, the text prompt is \"an underwater scene featuring a Black Marlin swimming swiftly, displaying its magnificent streamlined form and the vibrant colors of the ocean\",\nall the frames of video are as follows:\n\n<image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> " }, { "from": "gpt", "value": "visual quality: 3\n temporal consistency: 2\n dynamic degree: 3\n text-to-video alignment: 3\n factual consistency: 3\n" } ]

[ "1001247_00.jpg", "1001247_01.jpg", "1001247_02.jpg", "1001247_03.jpg", "1001247_04.jpg", "1001247_05.jpg", "1001247_06.jpg", "1001247_07.jpg", "1001247_08.jpg", "1001247_09.jpg", "1001247_10.jpg", "1001247_11.jpg", "1001247_12.jpg", "1001247_13.jpg", "1001247_14.jpg", "1001247_15.jpg" ]

[ { "from": "human", "value": "\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos,\nplease watch the following frames of a given video and see the text prompt for generating the video,\nthen give scores from 5 different dimensions:\n(1) visual quality: the quality of the video in terms of clearness, resolution, brightness, and color\n(2) temporal consistency, both the consistency of objects or humans and the smoothness of motion or movements\n(3) dynamic degree, the degree of dynamic changes\n(4) text-to-video alignment, the alignment between the text prompt and the video content\n(5) factual consistency, the consistency of the video content with the common-sense and factual knowledge\n\nfor each dimension, output a float number from 1.0 to 4.0,\nthe higher the number is, the better the video performs in that sub-score, \nthe lowest 1.0 means Bad, the highest 4.0 means Perfect/Real (the video is like a real video)\nHere is an output example:\nvisual quality: 3.2\ntemporal consistency: 2.7\ndynamic degree: 4.0\ntext-to-video alignment: 2.3\nfactual consistency: 1.8\n\nFor this video, the text prompt is \"Lead the trio into the Crystal Caves, a dazzling underground world filled with sparkling crystals. 4k HD Cinematic scene.\",\nall the frames of video are as follows:\n\n<image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> " }, { "from": "gpt", "value": "visual quality: 1\n temporal consistency: 1\n dynamic degree: 3\n text-to-video alignment: 2\n factual consistency: 1\n" } ]

[ "4001764_00.jpg", "4001764_01.jpg", "4001764_02.jpg", "4001764_03.jpg", "4001764_04.jpg", "4001764_05.jpg", "4001764_06.jpg", "4001764_07.jpg", "4001764_08.jpg", "4001764_09.jpg", "4001764_10.jpg", "4001764_11.jpg", "4001764_12.jpg", "4001764_13.jpg", "4001764_14.jpg", "4001764_15.jpg" ]

[ { "from": "human", "value": "\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos,\nplease watch the following frames of a given video and see the text prompt for generating the video,\nthen give scores from 5 different dimensions:\n(1) visual quality: the quality of the video in terms of clearness, resolution, brightness, and color\n(2) temporal consistency, both the consistency of objects or humans and the smoothness of motion or movements\n(3) dynamic degree, the degree of dynamic changes\n(4) text-to-video alignment, the alignment between the text prompt and the video content\n(5) factual consistency, the consistency of the video content with the common-sense and factual knowledge\n\nfor each dimension, output a float number from 1.0 to 4.0,\nthe higher the number is, the better the video performs in that sub-score, \nthe lowest 1.0 means Bad, the highest 4.0 means Perfect/Real (the video is like a real video)\nHere is an output example:\nvisual quality: 3.2\ntemporal consistency: 2.7\ndynamic degree: 4.0\ntext-to-video alignment: 2.3\nfactual consistency: 1.8\n\nFor this video, the text prompt is \"a steampunk fire truck, fired out from a volcano hiperrealism, detailed, high resolution\",\nall the frames of video are as follows:\n\n<image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> " }, { "from": "gpt", "value": "visual quality: 3\n temporal consistency: 2\n dynamic degree: 3\n text-to-video alignment: 3\n factual consistency: 3\n" } ]

[ "1000281_00.jpg", "1000281_01.jpg", "1000281_02.jpg", "1000281_03.jpg", "1000281_04.jpg", "1000281_05.jpg", "1000281_06.jpg", "1000281_07.jpg", "1000281_08.jpg", "1000281_09.jpg", "1000281_10.jpg", "1000281_11.jpg", "1000281_12.jpg", "1000281_13.jpg", "1000281_14.jpg", "1000281_15.jpg" ]

[ { "from": "human", "value": "\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos,\nplease watch the following frames of a given video and see the text prompt for generating the video,\nthen give scores from 5 different dimensions:\n(1) visual quality: the quality of the video in terms of clearness, resolution, brightness, and color\n(2) temporal consistency, both the consistency of objects or humans and the smoothness of motion or movements\n(3) dynamic degree, the degree of dynamic changes\n(4) text-to-video alignment, the alignment between the text prompt and the video content\n(5) factual consistency, the consistency of the video content with the common-sense and factual knowledge\n\nfor each dimension, output a float number from 1.0 to 4.0,\nthe higher the number is, the better the video performs in that sub-score, \nthe lowest 1.0 means Bad, the highest 4.0 means Perfect/Real (the video is like a real video)\nHere is an output example:\nvisual quality: 3.2\ntemporal consistency: 2.7\ndynamic degree: 4.0\ntext-to-video alignment: 2.3\nfactual consistency: 1.8\n\nFor this video, the text prompt is \"a girl on white silk is swimming in deep sea, beautiful fishs swimming behind her in a cinematic shot.\",\nall the frames of video are as follows:\n\n<image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> " }, { "from": "gpt", "value": "visual quality: 1\n temporal consistency: 2\n dynamic degree: 3\n text-to-video alignment: 2\n factual consistency: 1\n" } ]

[ "7001581_00.jpg", "7001581_01.jpg", "7001581_02.jpg", "7001581_03.jpg", "7001581_04.jpg", "7001581_05.jpg", "7001581_06.jpg", "7001581_07.jpg" ]

[ { "from": "human", "value": "\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos,\nplease watch the following frames of a given video and see the text prompt for generating the video,\nthen give scores from 5 different dimensions:\n(1) visual quality: the quality of the video in terms of clearness, resolution, brightness, and color\n(2) temporal consistency, both the consistency of objects or humans and the smoothness of motion or movements\n(3) dynamic degree, the degree of dynamic changes\n(4) text-to-video alignment, the alignment between the text prompt and the video content\n(5) factual consistency, the consistency of the video content with the common-sense and factual knowledge\n\nfor each dimension, output a float number from 1.0 to 4.0,\nthe higher the number is, the better the video performs in that sub-score, \nthe lowest 1.0 means Bad, the highest 4.0 means Perfect/Real (the video is like a real video)\nHere is an output example:\nvisual quality: 3.2\ntemporal consistency: 2.7\ndynamic degree: 4.0\ntext-to-video alignment: 2.3\nfactual consistency: 1.8\n\nFor this video, the text prompt is \"Invent a face veiled in shadows, with eyes that gleam with an enigmatic intensity. Add intricate, dark markings that flow along the contours of the face, suggesting a connection to ancient and mysterious forces\",\nall the frames of video are as follows:\n\n<image> <image> <image> <image> <image> <image> <image> <image> " }, { "from": "gpt", "value": "visual quality: 3\n temporal consistency: 3\n dynamic degree: 2\n text-to-video alignment: 3\n factual consistency: 3\n" } ]

[ "3001855_00.jpg", "3001855_01.jpg", "3001855_02.jpg", "3001855_03.jpg", "3001855_04.jpg", "3001855_05.jpg", "3001855_06.jpg", "3001855_07.jpg", "3001855_08.jpg", "3001855_09.jpg", "3001855_10.jpg", "3001855_11.jpg", "3001855_12.jpg", "3001855_13.jpg", "3001855_14.jpg", "3001855_15.jpg" ]

[ { "from": "human", "value": "\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos,\nplease watch the following frames of a given video and see the text prompt for generating the video,\nthen give scores from 5 different dimensions:\n(1) visual quality: the quality of the video in terms of clearness, resolution, brightness, and color\n(2) temporal consistency, both the consistency of objects or humans and the smoothness of motion or movements\n(3) dynamic degree, the degree of dynamic changes\n(4) text-to-video alignment, the alignment between the text prompt and the video content\n(5) factual consistency, the consistency of the video content with the common-sense and factual knowledge\n\nfor each dimension, output a float number from 1.0 to 4.0,\nthe higher the number is, the better the video performs in that sub-score, \nthe lowest 1.0 means Bad, the highest 4.0 means Perfect/Real (the video is like a real video)\nHere is an output example:\nvisual quality: 3.2\ntemporal consistency: 2.7\ndynamic degree: 4.0\ntext-to-video alignment: 2.3\nfactual consistency: 1.8\n\nFor this video, the text prompt is \"wind blow, waving sea, demon laugh\",\nall the frames of video are as follows:\n\n<image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> " }, { "from": "gpt", "value": "visual quality: 3\n temporal consistency: 2\n dynamic degree: 3\n text-to-video alignment: 2\n factual consistency: 3\n" } ]

[ "4000848_00.jpg", "4000848_01.jpg", "4000848_02.jpg", "4000848_03.jpg", "4000848_04.jpg", "4000848_05.jpg", "4000848_06.jpg", "4000848_07.jpg", "4000848_08.jpg", "4000848_09.jpg", "4000848_10.jpg", "4000848_11.jpg", "4000848_12.jpg", "4000848_13.jpg", "4000848_14.jpg", "4000848_15.jpg" ]

[ { "from": "human", "value": "\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos,\nplease watch the following frames of a given video and see the text prompt for generating the video,\nthen give scores from 5 different dimensions:\n(1) visual quality: the quality of the video in terms of clearness, resolution, brightness, and color\n(2) temporal consistency, both the consistency of objects or humans and the smoothness of motion or movements\n(3) dynamic degree, the degree of dynamic changes\n(4) text-to-video alignment, the alignment between the text prompt and the video content\n(5) factual consistency, the consistency of the video content with the common-sense and factual knowledge\n\nfor each dimension, output a float number from 1.0 to 4.0,\nthe higher the number is, the better the video performs in that sub-score, \nthe lowest 1.0 means Bad, the highest 4.0 means Perfect/Real (the video is like a real video)\nHere is an output example:\nvisual quality: 3.2\ntemporal consistency: 2.7\ndynamic degree: 4.0\ntext-to-video alignment: 2.3\nfactual consistency: 1.8\n\nFor this video, the text prompt is \"a cytpic box on a table with light shining on it, realistic\",\nall the frames of video are as follows:\n\n<image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> " }, { "from": "gpt", "value": "visual quality: 3\n temporal consistency: 2\n dynamic degree: 3\n text-to-video alignment: 3\n factual consistency: 3\n" } ]

[ "4001892_00.jpg", "4001892_01.jpg", "4001892_02.jpg", "4001892_03.jpg", "4001892_04.jpg", "4001892_05.jpg", "4001892_06.jpg", "4001892_07.jpg", "4001892_08.jpg", "4001892_09.jpg", "4001892_10.jpg", "4001892_11.jpg", "4001892_12.jpg", "4001892_13.jpg", "4001892_14.jpg", "4001892_15.jpg" ]

[ { "from": "human", "value": "\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos,\nplease watch the following frames of a given video and see the text prompt for generating the video,\nthen give scores from 5 different dimensions:\n(1) visual quality: the quality of the video in terms of clearness, resolution, brightness, and color\n(2) temporal consistency, both the consistency of objects or humans and the smoothness of motion or movements\n(3) dynamic degree, the degree of dynamic changes\n(4) text-to-video alignment, the alignment between the text prompt and the video content\n(5) factual consistency, the consistency of the video content with the common-sense and factual knowledge\n\nfor each dimension, output a float number from 1.0 to 4.0,\nthe higher the number is, the better the video performs in that sub-score, \nthe lowest 1.0 means Bad, the highest 4.0 means Perfect/Real (the video is like a real video)\nHere is an output example:\nvisual quality: 3.2\ntemporal consistency: 2.7\ndynamic degree: 4.0\ntext-to-video alignment: 2.3\nfactual consistency: 1.8\n\nFor this video, the text prompt is \"super realist horse running across an ethereal plane\",\nall the frames of video are as follows:\n\n<image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> " }, { "from": "gpt", "value": "visual quality: 3\n temporal consistency: 2\n dynamic degree: 3\n text-to-video alignment: 3\n factual consistency: 3\n" } ]

[ "7001450_00.jpg", "7001450_01.jpg", "7001450_02.jpg", "7001450_03.jpg", "7001450_04.jpg", "7001450_05.jpg", "7001450_06.jpg", "7001450_07.jpg" ]

[ { "from": "human", "value": "\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos,\nplease watch the following frames of a given video and see the text prompt for generating the video,\nthen give scores from 5 different dimensions:\n(1) visual quality: the quality of the video in terms of clearness, resolution, brightness, and color\n(2) temporal consistency, both the consistency of objects or humans and the smoothness of motion or movements\n(3) dynamic degree, the degree of dynamic changes\n(4) text-to-video alignment, the alignment between the text prompt and the video content\n(5) factual consistency, the consistency of the video content with the common-sense and factual knowledge\n\nfor each dimension, output a float number from 1.0 to 4.0,\nthe higher the number is, the better the video performs in that sub-score, \nthe lowest 1.0 means Bad, the highest 4.0 means Perfect/Real (the video is like a real video)\nHere is an output example:\nvisual quality: 3.2\ntemporal consistency: 2.7\ndynamic degree: 4.0\ntext-to-video alignment: 2.3\nfactual consistency: 1.8\n\nFor this video, the text prompt is \"Fire is burning everywhere and people are afraid\",\nall the frames of video are as follows:\n\n<image> <image> <image> <image> <image> <image> <image> <image> " }, { "from": "gpt", "value": "visual quality: 3\n temporal consistency: 3\n dynamic degree: 3\n text-to-video alignment: 2\n factual consistency: 3\n" } ]

[ "1002005_00.jpg", "1002005_01.jpg", "1002005_02.jpg", "1002005_03.jpg", "1002005_04.jpg", "1002005_05.jpg", "1002005_06.jpg", "1002005_07.jpg", "1002005_08.jpg", "1002005_09.jpg", "1002005_10.jpg", "1002005_11.jpg", "1002005_12.jpg", "1002005_13.jpg", "1002005_14.jpg", "1002005_15.jpg" ]

[ { "from": "human", "value": "\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos,\nplease watch the following frames of a given video and see the text prompt for generating the video,\nthen give scores from 5 different dimensions:\n(1) visual quality: the quality of the video in terms of clearness, resolution, brightness, and color\n(2) temporal consistency, both the consistency of objects or humans and the smoothness of motion or movements\n(3) dynamic degree, the degree of dynamic changes\n(4) text-to-video alignment, the alignment between the text prompt and the video content\n(5) factual consistency, the consistency of the video content with the common-sense and factual knowledge\n\nfor each dimension, output a float number from 1.0 to 4.0,\nthe higher the number is, the better the video performs in that sub-score, \nthe lowest 1.0 means Bad, the highest 4.0 means Perfect/Real (the video is like a real video)\nHere is an output example:\nvisual quality: 3.2\ntemporal consistency: 2.7\ndynamic degree: 4.0\ntext-to-video alignment: 2.3\nfactual consistency: 1.8\n\nFor this video, the text prompt is \"Girls wearing white silk JK short skirts dancing\",\nall the frames of video are as follows:\n\n<image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> " }, { "from": "gpt", "value": "visual quality: 1\n temporal consistency: 1\n dynamic degree: 3\n text-to-video alignment: 2\n factual consistency: 1\n" } ]

[ "5000472_00.jpg", "5000472_01.jpg", "5000472_02.jpg", "5000472_03.jpg", "5000472_04.jpg", "5000472_05.jpg", "5000472_06.jpg", "5000472_07.jpg", "5000472_08.jpg", "5000472_09.jpg", "5000472_10.jpg", "5000472_11.jpg", "5000472_12.jpg", "5000472_13.jpg", "5000472_14.jpg", "5000472_15.jpg" ]

[ { "from": "human", "value": "\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos,\nplease watch the following frames of a given video and see the text prompt for generating the video,\nthen give scores from 5 different dimensions:\n(1) visual quality: the quality of the video in terms of clearness, resolution, brightness, and color\n(2) temporal consistency, both the consistency of objects or humans and the smoothness of motion or movements\n(3) dynamic degree, the degree of dynamic changes\n(4) text-to-video alignment, the alignment between the text prompt and the video content\n(5) factual consistency, the consistency of the video content with the common-sense and factual knowledge\n\nfor each dimension, output a float number from 1.0 to 4.0,\nthe higher the number is, the better the video performs in that sub-score, \nthe lowest 1.0 means Bad, the highest 4.0 means Perfect/Real (the video is like a real video)\nHere is an output example:\nvisual quality: 3.2\ntemporal consistency: 2.7\ndynamic degree: 4.0\ntext-to-video alignment: 2.3\nfactual consistency: 1.8\n\nFor this video, the text prompt is \"A mirror in a room is breaking and splattering\",\nall the frames of video are as follows:\n\n<image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> " }, { "from": "gpt", "value": "visual quality: 3\n temporal consistency: 1\n dynamic degree: 3\n text-to-video alignment: 2\n factual consistency: 1\n" } ]

[ "1002623_00.jpg", "1002623_01.jpg", "1002623_02.jpg", "1002623_03.jpg", "1002623_04.jpg", "1002623_05.jpg", "1002623_06.jpg", "1002623_07.jpg", "1002623_08.jpg", "1002623_09.jpg", "1002623_10.jpg", "1002623_11.jpg", "1002623_12.jpg", "1002623_13.jpg", "1002623_14.jpg", "1002623_15.jpg" ]

[ { "from": "human", "value": "\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos,\nplease watch the following frames of a given video and see the text prompt for generating the video,\nthen give scores from 5 different dimensions:\n(1) visual quality: the quality of the video in terms of clearness, resolution, brightness, and color\n(2) temporal consistency, both the consistency of objects or humans and the smoothness of motion or movements\n(3) dynamic degree, the degree of dynamic changes\n(4) text-to-video alignment, the alignment between the text prompt and the video content\n(5) factual consistency, the consistency of the video content with the common-sense and factual knowledge\n\nfor each dimension, output a float number from 1.0 to 4.0,\nthe higher the number is, the better the video performs in that sub-score, \nthe lowest 1.0 means Bad, the highest 4.0 means Perfect/Real (the video is like a real video)\nHere is an output example:\nvisual quality: 3.2\ntemporal consistency: 2.7\ndynamic degree: 4.0\ntext-to-video alignment: 2.3\nfactual consistency: 1.8\n\nFor this video, the text prompt is \"A dog with brown hair in a Disney Pixar movie style.\",\nall the frames of video are as follows:\n\n<image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> " }, { "from": "gpt", "value": "visual quality: 1\n temporal consistency: 2\n dynamic degree: 3\n text-to-video alignment: 2\n factual consistency: 1\n" } ]

[ "0001942_00.jpg", "0001942_01.jpg", "0001942_02.jpg", "0001942_03.jpg", "0001942_04.jpg", "0001942_05.jpg", "0001942_06.jpg", "0001942_07.jpg", "0001942_08.jpg", "0001942_09.jpg", "0001942_10.jpg", "0001942_11.jpg", "0001942_12.jpg", "0001942_13.jpg", "0001942_14.jpg", "0001942_15.jpg", "0001942_16.jpg", "0001942_17.jpg", "0001942_18.jpg", "0001942_19.jpg", "0001942_20.jpg", "0001942_21.jpg", "0001942_22.jpg", "0001942_23.jpg" ]

[ { "from": "human", "value": "\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos,\nplease watch the following frames of a given video and see the text prompt for generating the video,\nthen give scores from 5 different dimensions:\n(1) visual quality: the quality of the video in terms of clearness, resolution, brightness, and color\n(2) temporal consistency, both the consistency of objects or humans and the smoothness of motion or movements\n(3) dynamic degree, the degree of dynamic changes\n(4) text-to-video alignment, the alignment between the text prompt and the video content\n(5) factual consistency, the consistency of the video content with the common-sense and factual knowledge\n\nfor each dimension, output a float number from 1.0 to 4.0,\nthe higher the number is, the better the video performs in that sub-score, \nthe lowest 1.0 means Bad, the highest 4.0 means Perfect/Real (the video is like a real video)\nHere is an output example:\nvisual quality: 3.2\ntemporal consistency: 2.7\ndynamic degree: 4.0\ntext-to-video alignment: 2.3\nfactual consistency: 1.8\n\nFor this video, the text prompt is \"abraham lincoln in a fist fight with uncle sam while the the dallar cheerleaders cheer for them in the background.\",\nall the frames of video are as follows:\n\n<image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> " }, { "from": "gpt", "value": "visual quality: 2\n temporal consistency: 2\n dynamic degree: 3\n text-to-video alignment: 3\n factual consistency: 3\n" } ]

[ "6000346_00.jpg", "6000346_01.jpg", "6000346_02.jpg", "6000346_03.jpg", "6000346_04.jpg", "6000346_05.jpg", "6000346_06.jpg", "6000346_07.jpg", "6000346_08.jpg", "6000346_09.jpg", "6000346_10.jpg", "6000346_11.jpg", "6000346_12.jpg", "6000346_13.jpg", "6000346_14.jpg", "6000346_15.jpg" ]

[ { "from": "human", "value": "\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos,\nplease watch the following frames of a given video and see the text prompt for generating the video,\nthen give scores from 5 different dimensions:\n(1) visual quality: the quality of the video in terms of clearness, resolution, brightness, and color\n(2) temporal consistency, both the consistency of objects or humans and the smoothness of motion or movements\n(3) dynamic degree, the degree of dynamic changes\n(4) text-to-video alignment, the alignment between the text prompt and the video content\n(5) factual consistency, the consistency of the video content with the common-sense and factual knowledge\n\nfor each dimension, output a float number from 1.0 to 4.0,\nthe higher the number is, the better the video performs in that sub-score, \nthe lowest 1.0 means Bad, the highest 4.0 means Perfect/Real (the video is like a real video)\nHere is an output example:\nvisual quality: 3.2\ntemporal consistency: 2.7\ndynamic degree: 4.0\ntext-to-video alignment: 2.3\nfactual consistency: 1.8\n\nFor this video, the text prompt is \"The sun sets over a tranquil seaside town. The waves gently lap against the shore as seagulls soar overhead. On the bustling boardwalk, children laugh and play, and a street musician strums a melodic tune.\",\nall the frames of video are as follows:\n\n<image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> " }, { "from": "gpt", "value": "visual quality: 1\n temporal consistency: 2\n dynamic degree: 3\n text-to-video alignment: 1\n factual consistency: 1\n" } ]

[ "4001087_00.jpg", "4001087_01.jpg", "4001087_02.jpg", "4001087_03.jpg", "4001087_04.jpg", "4001087_05.jpg", "4001087_06.jpg", "4001087_07.jpg", "4001087_08.jpg", "4001087_09.jpg", "4001087_10.jpg", "4001087_11.jpg", "4001087_12.jpg", "4001087_13.jpg", "4001087_14.jpg", "4001087_15.jpg" ]

[ { "from": "human", "value": "\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos,\nplease watch the following frames of a given video and see the text prompt for generating the video,\nthen give scores from 5 different dimensions:\n(1) visual quality: the quality of the video in terms of clearness, resolution, brightness, and color\n(2) temporal consistency, both the consistency of objects or humans and the smoothness of motion or movements\n(3) dynamic degree, the degree of dynamic changes\n(4) text-to-video alignment, the alignment between the text prompt and the video content\n(5) factual consistency, the consistency of the video content with the common-sense and factual knowledge\n\nfor each dimension, output a float number from 1.0 to 4.0,\nthe higher the number is, the better the video performs in that sub-score, \nthe lowest 1.0 means Bad, the highest 4.0 means Perfect/Real (the video is like a real video)\nHere is an output example:\nvisual quality: 3.2\ntemporal consistency: 2.7\ndynamic degree: 4.0\ntext-to-video alignment: 2.3\nfactual consistency: 1.8\n\nFor this video, the text prompt is \"faceless man in a dark, liminal space\",\nall the frames of video are as follows:\n\n<image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> " }, { "from": "gpt", "value": "visual quality: 2\n temporal consistency: 3\n dynamic degree: 3\n text-to-video alignment: 2\n factual consistency: 3\n" } ]

[ "3001311_00.jpg", "3001311_01.jpg", "3001311_02.jpg", "3001311_03.jpg", "3001311_04.jpg", "3001311_05.jpg", "3001311_06.jpg", "3001311_07.jpg", "3001311_08.jpg", "3001311_09.jpg", "3001311_10.jpg", "3001311_11.jpg", "3001311_12.jpg", "3001311_13.jpg", "3001311_14.jpg", "3001311_15.jpg" ]

[ { "from": "human", "value": "\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos,\nplease watch the following frames of a given video and see the text prompt for generating the video,\nthen give scores from 5 different dimensions:\n(1) visual quality: the quality of the video in terms of clearness, resolution, brightness, and color\n(2) temporal consistency, both the consistency of objects or humans and the smoothness of motion or movements\n(3) dynamic degree, the degree of dynamic changes\n(4) text-to-video alignment, the alignment between the text prompt and the video content\n(5) factual consistency, the consistency of the video content with the common-sense and factual knowledge\n\nfor each dimension, output a float number from 1.0 to 4.0,\nthe higher the number is, the better the video performs in that sub-score, \nthe lowest 1.0 means Bad, the highest 4.0 means Perfect/Real (the video is like a real video)\nHere is an output example:\nvisual quality: 3.2\ntemporal consistency: 2.7\ndynamic degree: 4.0\ntext-to-video alignment: 2.3\nfactual consistency: 1.8\n\nFor this video, the text prompt is \"two rival groups of animals lived, the Wolf Pack and the Fox Clan.\",\nall the frames of video are as follows:\n\n<image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> " }, { "from": "gpt", "value": "visual quality: 2\n temporal consistency: 1\n dynamic degree: 3\n text-to-video alignment: 1\n factual consistency: 1\n" } ]

[ "7001242_00.jpg", "7001242_01.jpg", "7001242_02.jpg", "7001242_03.jpg", "7001242_04.jpg", "7001242_05.jpg", "7001242_06.jpg", "7001242_07.jpg" ]

[ { "from": "human", "value": "\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos,\nplease watch the following frames of a given video and see the text prompt for generating the video,\nthen give scores from 5 different dimensions:\n(1) visual quality: the quality of the video in terms of clearness, resolution, brightness, and color\n(2) temporal consistency, both the consistency of objects or humans and the smoothness of motion or movements\n(3) dynamic degree, the degree of dynamic changes\n(4) text-to-video alignment, the alignment between the text prompt and the video content\n(5) factual consistency, the consistency of the video content with the common-sense and factual knowledge\n\nfor each dimension, output a float number from 1.0 to 4.0,\nthe higher the number is, the better the video performs in that sub-score, \nthe lowest 1.0 means Bad, the highest 4.0 means Perfect/Real (the video is like a real video)\nHere is an output example:\nvisual quality: 3.2\ntemporal consistency: 2.7\ndynamic degree: 4.0\ntext-to-video alignment: 2.3\nfactual consistency: 1.8\n\nFor this video, the text prompt is \"Generate a captivating image of a young Pakistani girl, Ayesha, standing in awe under a stormy sky. Convey the mystery and energy of the moment as bolts of lightning dance in the distance, reflecting the curiosity and innocence of a child discovering her extraordinary connection to the elements.\",\nall the frames of video are as follows:\n\n<image> <image> <image> <image> <image> <image> <image> <image> " }, { "from": "gpt", "value": "visual quality: 2\n temporal consistency: 2\n dynamic degree: 3\n text-to-video alignment: 3\n factual consistency: 3\n" } ]

[ "2002669_00.jpg", "2002669_01.jpg", "2002669_02.jpg", "2002669_03.jpg", "2002669_04.jpg", "2002669_05.jpg", "2002669_06.jpg", "2002669_07.jpg", "2002669_08.jpg", "2002669_09.jpg", "2002669_10.jpg", "2002669_11.jpg", "2002669_12.jpg", "2002669_13.jpg", "2002669_14.jpg", "2002669_15.jpg" ]

[ { "from": "human", "value": "\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos,\nplease watch the following frames of a given video and see the text prompt for generating the video,\nthen give scores from 5 different dimensions:\n(1) visual quality: the quality of the video in terms of clearness, resolution, brightness, and color\n(2) temporal consistency, both the consistency of objects or humans and the smoothness of motion or movements\n(3) dynamic degree, the degree of dynamic changes\n(4) text-to-video alignment, the alignment between the text prompt and the video content\n(5) factual consistency, the consistency of the video content with the common-sense and factual knowledge\n\nfor each dimension, output a float number from 1.0 to 4.0,\nthe higher the number is, the better the video performs in that sub-score, \nthe lowest 1.0 means Bad, the highest 4.0 means Perfect/Real (the video is like a real video)\nHere is an output example:\nvisual quality: 3.2\ntemporal consistency: 2.7\ndynamic degree: 4.0\ntext-to-video alignment: 2.3\nfactual consistency: 1.8\n\nFor this video, the text prompt is \"3d stylized little game building library, intrincade details, cinematic, octane render\",\nall the frames of video are as follows:\n\n<image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> " }, { "from": "gpt", "value": "visual quality: 2\n temporal consistency: 3\n dynamic degree: 2\n text-to-video alignment: 3\n factual consistency: 3\n" } ]

[ "3002675_00.jpg", "3002675_01.jpg", "3002675_02.jpg", "3002675_03.jpg", "3002675_04.jpg", "3002675_05.jpg", "3002675_06.jpg", "3002675_07.jpg", "3002675_08.jpg", "3002675_09.jpg", "3002675_10.jpg", "3002675_11.jpg", "3002675_12.jpg", "3002675_13.jpg", "3002675_14.jpg", "3002675_15.jpg" ]

[ { "from": "human", "value": "\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos,\nplease watch the following frames of a given video and see the text prompt for generating the video,\nthen give scores from 5 different dimensions:\n(1) visual quality: the quality of the video in terms of clearness, resolution, brightness, and color\n(2) temporal consistency, both the consistency of objects or humans and the smoothness of motion or movements\n(3) dynamic degree, the degree of dynamic changes\n(4) text-to-video alignment, the alignment between the text prompt and the video content\n(5) factual consistency, the consistency of the video content with the common-sense and factual knowledge\n\nfor each dimension, output a float number from 1.0 to 4.0,\nthe higher the number is, the better the video performs in that sub-score, \nthe lowest 1.0 means Bad, the highest 4.0 means Perfect/Real (the video is like a real video)\nHere is an output example:\nvisual quality: 3.2\ntemporal consistency: 2.7\ndynamic degree: 4.0\ntext-to-video alignment: 2.3\nfactual consistency: 1.8\n\nFor this video, the text prompt is \"A dog, computer working , captured in 4K ,\",\nall the frames of video are as follows:\n\n<image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> " }, { "from": "gpt", "value": "visual quality: 2\n temporal consistency: 3\n dynamic degree: 3\n text-to-video alignment: 3\n factual consistency: 3\n" } ]

[ "0002070_00.jpg", "0002070_01.jpg", "0002070_02.jpg", "0002070_03.jpg", "0002070_04.jpg", "0002070_05.jpg", "0002070_06.jpg", "0002070_07.jpg", "0002070_08.jpg", "0002070_09.jpg", "0002070_10.jpg", "0002070_11.jpg", "0002070_12.jpg", "0002070_13.jpg", "0002070_14.jpg", "0002070_15.jpg", "0002070_16.jpg", "0002070_17.jpg", "0002070_18.jpg", "0002070_19.jpg", "0002070_20.jpg", "0002070_21.jpg", "0002070_22.jpg", "0002070_23.jpg" ]

[ { "from": "human", "value": "\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos,\nplease watch the following frames of a given video and see the text prompt for generating the video,\nthen give scores from 5 different dimensions:\n(1) visual quality: the quality of the video in terms of clearness, resolution, brightness, and color\n(2) temporal consistency, both the consistency of objects or humans and the smoothness of motion or movements\n(3) dynamic degree, the degree of dynamic changes\n(4) text-to-video alignment, the alignment between the text prompt and the video content\n(5) factual consistency, the consistency of the video content with the common-sense and factual knowledge\n\nfor each dimension, output a float number from 1.0 to 4.0,\nthe higher the number is, the better the video performs in that sub-score, \nthe lowest 1.0 means Bad, the highest 4.0 means Perfect/Real (the video is like a real video)\nHere is an output example:\nvisual quality: 3.2\ntemporal consistency: 2.7\ndynamic degree: 4.0\ntext-to-video alignment: 2.3\nfactual consistency: 1.8\n\nFor this video, the text prompt is \"one cool dude wearing leather jacket and sunglasses driving harley davidson motorbike on the highway, high speed, driving fast, motionblur\",\nall the frames of video are as follows:\n\n<image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> " }, { "from": "gpt", "value": "visual quality: 3\n temporal consistency: 3\n dynamic degree: 2\n text-to-video alignment: 3\n factual consistency: 3\n" } ]

[ "3002247_00.jpg", "3002247_01.jpg", "3002247_02.jpg", "3002247_03.jpg", "3002247_04.jpg", "3002247_05.jpg", "3002247_06.jpg", "3002247_07.jpg", "3002247_08.jpg", "3002247_09.jpg", "3002247_10.jpg", "3002247_11.jpg", "3002247_12.jpg", "3002247_13.jpg", "3002247_14.jpg", "3002247_15.jpg" ]

[ { "from": "human", "value": "\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos,\nplease watch the following frames of a given video and see the text prompt for generating the video,\nthen give scores from 5 different dimensions:\n(1) visual quality: the quality of the video in terms of clearness, resolution, brightness, and color\n(2) temporal consistency, both the consistency of objects or humans and the smoothness of motion or movements\n(3) dynamic degree, the degree of dynamic changes\n(4) text-to-video alignment, the alignment between the text prompt and the video content\n(5) factual consistency, the consistency of the video content with the common-sense and factual knowledge\n\nfor each dimension, output a float number from 1.0 to 4.0,\nthe higher the number is, the better the video performs in that sub-score, \nthe lowest 1.0 means Bad, the highest 4.0 means Perfect/Real (the video is like a real video)\nHere is an output example:\nvisual quality: 3.2\ntemporal consistency: 2.7\ndynamic degree: 4.0\ntext-to-video alignment: 2.3\nfactual consistency: 1.8\n\nFor this video, the text prompt is \"Cyrus the great fighting Romans near of Persepolis palace\",\nall the frames of video are as follows:\n\n<image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> " }, { "from": "gpt", "value": "visual quality: 3\n temporal consistency: 2\n dynamic degree: 3\n text-to-video alignment: 2\n factual consistency: 3\n" } ]

[ "7000546_00.jpg", "7000546_01.jpg", "7000546_02.jpg", "7000546_03.jpg", "7000546_04.jpg", "7000546_05.jpg", "7000546_06.jpg", "7000546_07.jpg" ]

[ { "from": "human", "value": "\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos,\nplease watch the following frames of a given video and see the text prompt for generating the video,\nthen give scores from 5 different dimensions:\n(1) visual quality: the quality of the video in terms of clearness, resolution, brightness, and color\n(2) temporal consistency, both the consistency of objects or humans and the smoothness of motion or movements\n(3) dynamic degree, the degree of dynamic changes\n(4) text-to-video alignment, the alignment between the text prompt and the video content\n(5) factual consistency, the consistency of the video content with the common-sense and factual knowledge\n\nfor each dimension, output a float number from 1.0 to 4.0,\nthe higher the number is, the better the video performs in that sub-score, \nthe lowest 1.0 means Bad, the highest 4.0 means Perfect/Real (the video is like a real video)\nHere is an output example:\nvisual quality: 3.2\ntemporal consistency: 2.7\ndynamic degree: 4.0\ntext-to-video alignment: 2.3\nfactual consistency: 1.8\n\nFor this video, the text prompt is \"2d style animation of multiple people walking in a crowded,busy sidewalk,cars driving on city street.\",\nall the frames of video are as follows:\n\n<image> <image> <image> <image> <image> <image> <image> <image> " }, { "from": "gpt", "value": "visual quality: 1\n temporal consistency: 1\n dynamic degree: 3\n text-to-video alignment: 3\n factual consistency: 3\n" } ]

[ "6000644_00.jpg", "6000644_01.jpg", "6000644_02.jpg", "6000644_03.jpg", "6000644_04.jpg", "6000644_05.jpg", "6000644_06.jpg", "6000644_07.jpg", "6000644_08.jpg", "6000644_09.jpg", "6000644_10.jpg", "6000644_11.jpg", "6000644_12.jpg", "6000644_13.jpg", "6000644_14.jpg", "6000644_15.jpg" ]

[ { "from": "human", "value": "\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos,\nplease watch the following frames of a given video and see the text prompt for generating the video,\nthen give scores from 5 different dimensions:\n(1) visual quality: the quality of the video in terms of clearness, resolution, brightness, and color\n(2) temporal consistency, both the consistency of objects or humans and the smoothness of motion or movements\n(3) dynamic degree, the degree of dynamic changes\n(4) text-to-video alignment, the alignment between the text prompt and the video content\n(5) factual consistency, the consistency of the video content with the common-sense and factual knowledge\n\nfor each dimension, output a float number from 1.0 to 4.0,\nthe higher the number is, the better the video performs in that sub-score, \nthe lowest 1.0 means Bad, the highest 4.0 means Perfect/Real (the video is like a real video)\nHere is an output example:\nvisual quality: 3.2\ntemporal consistency: 2.7\ndynamic degree: 4.0\ntext-to-video alignment: 2.3\nfactual consistency: 1.8\n\nFor this video, the text prompt is \"captain America and spider man are fighting, captain throwing Strom bracker to Spidey but spider man caught the strom bracker with his web shooter.\",\nall the frames of video are as follows:\n\n<image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> " }, { "from": "gpt", "value": "visual quality: 1\n temporal consistency: 2\n dynamic degree: 3\n text-to-video alignment: 3\n factual consistency: 1\n" } ]

[ "1002265_00.jpg", "1002265_01.jpg", "1002265_02.jpg", "1002265_03.jpg", "1002265_04.jpg", "1002265_05.jpg", "1002265_06.jpg", "1002265_07.jpg", "1002265_08.jpg", "1002265_09.jpg", "1002265_10.jpg", "1002265_11.jpg", "1002265_12.jpg", "1002265_13.jpg", "1002265_14.jpg", "1002265_15.jpg" ]

[ { "from": "human", "value": "\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos,\nplease watch the following frames of a given video and see the text prompt for generating the video,\nthen give scores from 5 different dimensions:\n(1) visual quality: the quality of the video in terms of clearness, resolution, brightness, and color\n(2) temporal consistency, both the consistency of objects or humans and the smoothness of motion or movements\n(3) dynamic degree, the degree of dynamic changes\n(4) text-to-video alignment, the alignment between the text prompt and the video content\n(5) factual consistency, the consistency of the video content with the common-sense and factual knowledge\n\nfor each dimension, output a float number from 1.0 to 4.0,\nthe higher the number is, the better the video performs in that sub-score, \nthe lowest 1.0 means Bad, the highest 4.0 means Perfect/Real (the video is like a real video)\nHere is an output example:\nvisual quality: 3.2\ntemporal consistency: 2.7\ndynamic degree: 4.0\ntext-to-video alignment: 2.3\nfactual consistency: 1.8\n\nFor this video, the text prompt is \"cartoon of a music speaker playing loud music, rave flyer style, Keith Haring illustration, bold lines, photocopy, black and white, simple\",\nall the frames of video are as follows:\n\n<image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> " }, { "from": "gpt", "value": "visual quality: 1\n temporal consistency: 1\n dynamic degree: 3\n text-to-video alignment: 2\n factual consistency: 1\n" } ]

[ "6000441_00.jpg", "6000441_01.jpg", "6000441_02.jpg", "6000441_03.jpg", "6000441_04.jpg", "6000441_05.jpg", "6000441_06.jpg", "6000441_07.jpg", "6000441_08.jpg", "6000441_09.jpg", "6000441_10.jpg", "6000441_11.jpg", "6000441_12.jpg", "6000441_13.jpg", "6000441_14.jpg", "6000441_15.jpg" ]

[ { "from": "human", "value": "\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos,\nplease watch the following frames of a given video and see the text prompt for generating the video,\nthen give scores from 5 different dimensions:\n(1) visual quality: the quality of the video in terms of clearness, resolution, brightness, and color\n(2) temporal consistency, both the consistency of objects or humans and the smoothness of motion or movements\n(3) dynamic degree, the degree of dynamic changes\n(4) text-to-video alignment, the alignment between the text prompt and the video content\n(5) factual consistency, the consistency of the video content with the common-sense and factual knowledge\n\nfor each dimension, output a float number from 1.0 to 4.0,\nthe higher the number is, the better the video performs in that sub-score, \nthe lowest 1.0 means Bad, the highest 4.0 means Perfect/Real (the video is like a real video)\nHere is an output example:\nvisual quality: 3.2\ntemporal consistency: 2.7\ndynamic degree: 4.0\ntext-to-video alignment: 2.3\nfactual consistency: 1.8\n\nFor this video, the text prompt is \"Four mermaids trying to drown a 13 y.o girl in a blue dress standing in a rock at the sea, 70s dark fantasy style, live action, realistic, cinematic, vintage, HD realistic, ULTRA HD, 70s movie style\",\nall the frames of video are as follows:\n\n<image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> " }, { "from": "gpt", "value": "visual quality: 3\n temporal consistency: 2\n dynamic degree: 3\n text-to-video alignment: 1\n factual consistency: 1\n" } ]

[ "5000446_00.jpg", "5000446_01.jpg", "5000446_02.jpg", "5000446_03.jpg", "5000446_04.jpg", "5000446_05.jpg", "5000446_06.jpg", "5000446_07.jpg", "5000446_08.jpg", "5000446_09.jpg", "5000446_10.jpg", "5000446_11.jpg", "5000446_12.jpg", "5000446_13.jpg", "5000446_14.jpg", "5000446_15.jpg" ]

[ { "from": "human", "value": "\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos,\nplease watch the following frames of a given video and see the text prompt for generating the video,\nthen give scores from 5 different dimensions:\n(1) visual quality: the quality of the video in terms of clearness, resolution, brightness, and color\n(2) temporal consistency, both the consistency of objects or humans and the smoothness of motion or movements\n(3) dynamic degree, the degree of dynamic changes\n(4) text-to-video alignment, the alignment between the text prompt and the video content\n(5) factual consistency, the consistency of the video content with the common-sense and factual knowledge\n\nfor each dimension, output a float number from 1.0 to 4.0,\nthe higher the number is, the better the video performs in that sub-score, \nthe lowest 1.0 means Bad, the highest 4.0 means Perfect/Real (the video is like a real video)\nHere is an output example:\nvisual quality: 3.2\ntemporal consistency: 2.7\ndynamic degree: 4.0\ntext-to-video alignment: 2.3\nfactual consistency: 1.8\n\nFor this video, the text prompt is \"alien waving on a remote island\",\nall the frames of video are as follows:\n\n<image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> " }, { "from": "gpt", "value": "visual quality: 3\n temporal consistency: 2\n dynamic degree: 3\n text-to-video alignment: 3\n factual consistency: 2\n" } ]

[ "4000297_00.jpg", "4000297_01.jpg", "4000297_02.jpg", "4000297_03.jpg", "4000297_04.jpg", "4000297_05.jpg", "4000297_06.jpg", "4000297_07.jpg", "4000297_08.jpg", "4000297_09.jpg", "4000297_10.jpg", "4000297_11.jpg", "4000297_12.jpg", "4000297_13.jpg", "4000297_14.jpg", "4000297_15.jpg" ]

[ { "from": "human", "value": "\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos,\nplease watch the following frames of a given video and see the text prompt for generating the video,\nthen give scores from 5 different dimensions:\n(1) visual quality: the quality of the video in terms of clearness, resolution, brightness, and color\n(2) temporal consistency, both the consistency of objects or humans and the smoothness of motion or movements\n(3) dynamic degree, the degree of dynamic changes\n(4) text-to-video alignment, the alignment between the text prompt and the video content\n(5) factual consistency, the consistency of the video content with the common-sense and factual knowledge\n\nfor each dimension, output a float number from 1.0 to 4.0,\nthe higher the number is, the better the video performs in that sub-score, \nthe lowest 1.0 means Bad, the highest 4.0 means Perfect/Real (the video is like a real video)\nHere is an output example:\nvisual quality: 3.2\ntemporal consistency: 2.7\ndynamic degree: 4.0\ntext-to-video alignment: 2.3\nfactual consistency: 1.8\n\nFor this video, the text prompt is \"create a flying pegasus in sky\",\nall the frames of video are as follows:\n\n<image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> " }, { "from": "gpt", "value": "visual quality: 3\n temporal consistency: 2\n dynamic degree: 3\n text-to-video alignment: 3\n factual consistency: 2\n" } ]

[ "2000361_00.jpg", "2000361_01.jpg", "2000361_02.jpg", "2000361_03.jpg", "2000361_04.jpg", "2000361_05.jpg", "2000361_06.jpg", "2000361_07.jpg", "2000361_08.jpg", "2000361_09.jpg", "2000361_10.jpg", "2000361_11.jpg", "2000361_12.jpg", "2000361_13.jpg", "2000361_14.jpg", "2000361_15.jpg" ]

[ { "from": "human", "value": "\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos,\nplease watch the following frames of a given video and see the text prompt for generating the video,\nthen give scores from 5 different dimensions:\n(1) visual quality: the quality of the video in terms of clearness, resolution, brightness, and color\n(2) temporal consistency, both the consistency of objects or humans and the smoothness of motion or movements\n(3) dynamic degree, the degree of dynamic changes\n(4) text-to-video alignment, the alignment between the text prompt and the video content\n(5) factual consistency, the consistency of the video content with the common-sense and factual knowledge\n\nfor each dimension, output a float number from 1.0 to 4.0,\nthe higher the number is, the better the video performs in that sub-score, \nthe lowest 1.0 means Bad, the highest 4.0 means Perfect/Real (the video is like a real video)\nHere is an output example:\nvisual quality: 3.2\ntemporal consistency: 2.7\ndynamic degree: 4.0\ntext-to-video alignment: 2.3\nfactual consistency: 1.8\n\nFor this video, the text prompt is \"Siren Head, Alfred Stieglitz grainy black and white 1900s\",\nall the frames of video are as follows:\n\n<image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> " }, { "from": "gpt", "value": "visual quality: 2\n temporal consistency: 3\n dynamic degree: 3\n text-to-video alignment: 2\n factual consistency: 1\n" } ]

[ "1000907_00.jpg", "1000907_01.jpg", "1000907_02.jpg", "1000907_03.jpg", "1000907_04.jpg", "1000907_05.jpg", "1000907_06.jpg", "1000907_07.jpg", "1000907_08.jpg", "1000907_09.jpg", "1000907_10.jpg", "1000907_11.jpg", "1000907_12.jpg", "1000907_13.jpg", "1000907_14.jpg", "1000907_15.jpg" ]

[ { "from": "human", "value": "\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos,\nplease watch the following frames of a given video and see the text prompt for generating the video,\nthen give scores from 5 different dimensions:\n(1) visual quality: the quality of the video in terms of clearness, resolution, brightness, and color\n(2) temporal consistency, both the consistency of objects or humans and the smoothness of motion or movements\n(3) dynamic degree, the degree of dynamic changes\n(4) text-to-video alignment, the alignment between the text prompt and the video content\n(5) factual consistency, the consistency of the video content with the common-sense and factual knowledge\n\nfor each dimension, output a float number from 1.0 to 4.0,\nthe higher the number is, the better the video performs in that sub-score, \nthe lowest 1.0 means Bad, the highest 4.0 means Perfect/Real (the video is like a real video)\nHere is an output example:\nvisual quality: 3.2\ntemporal consistency: 2.7\ndynamic degree: 4.0\ntext-to-video alignment: 2.3\nfactual consistency: 1.8\n\nFor this video, the text prompt is \"Sylvester Stallone box fighting with others\",\nall the frames of video are as follows:\n\n<image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> " }, { "from": "gpt", "value": "visual quality: 1\n temporal consistency: 2\n dynamic degree: 3\n text-to-video alignment: 2\n factual consistency: 1\n" } ]

[ "3000616_00.jpg", "3000616_01.jpg", "3000616_02.jpg", "3000616_03.jpg", "3000616_04.jpg", "3000616_05.jpg", "3000616_06.jpg", "3000616_07.jpg", "3000616_08.jpg", "3000616_09.jpg", "3000616_10.jpg", "3000616_11.jpg", "3000616_12.jpg", "3000616_13.jpg", "3000616_14.jpg", "3000616_15.jpg" ]

[ { "from": "human", "value": "\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos,\nplease watch the following frames of a given video and see the text prompt for generating the video,\nthen give scores from 5 different dimensions:\n(1) visual quality: the quality of the video in terms of clearness, resolution, brightness, and color\n(2) temporal consistency, both the consistency of objects or humans and the smoothness of motion or movements\n(3) dynamic degree, the degree of dynamic changes\n(4) text-to-video alignment, the alignment between the text prompt and the video content\n(5) factual consistency, the consistency of the video content with the common-sense and factual knowledge\n\nfor each dimension, output a float number from 1.0 to 4.0,\nthe higher the number is, the better the video performs in that sub-score, \nthe lowest 1.0 means Bad, the highest 4.0 means Perfect/Real (the video is like a real video)\nHere is an output example:\nvisual quality: 3.2\ntemporal consistency: 2.7\ndynamic degree: 4.0\ntext-to-video alignment: 2.3\nfactual consistency: 1.8\n\nFor this video, the text prompt is \"a cowboy saying hello waving hands eye contact with evil intentions\",\nall the frames of video are as follows:\n\n<image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> " }, { "from": "gpt", "value": "visual quality: 2\n temporal consistency: 3\n dynamic degree: 3\n text-to-video alignment: 3\n factual consistency: 2\n" } ]

[ "3000693_00.jpg", "3000693_01.jpg", "3000693_02.jpg", "3000693_03.jpg", "3000693_04.jpg", "3000693_05.jpg", "3000693_06.jpg", "3000693_07.jpg", "3000693_08.jpg", "3000693_09.jpg", "3000693_10.jpg", "3000693_11.jpg", "3000693_12.jpg", "3000693_13.jpg", "3000693_14.jpg", "3000693_15.jpg" ]

[ { "from": "human", "value": "\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos,\nplease watch the following frames of a given video and see the text prompt for generating the video,\nthen give scores from 5 different dimensions:\n(1) visual quality: the quality of the video in terms of clearness, resolution, brightness, and color\n(2) temporal consistency, both the consistency of objects or humans and the smoothness of motion or movements\n(3) dynamic degree, the degree of dynamic changes\n(4) text-to-video alignment, the alignment between the text prompt and the video content\n(5) factual consistency, the consistency of the video content with the common-sense and factual knowledge\n\nfor each dimension, output a float number from 1.0 to 4.0,\nthe higher the number is, the better the video performs in that sub-score, \nthe lowest 1.0 means Bad, the highest 4.0 means Perfect/Real (the video is like a real video)\nHere is an output example:\nvisual quality: 3.2\ntemporal consistency: 2.7\ndynamic degree: 4.0\ntext-to-video alignment: 2.3\nfactual consistency: 1.8\n\nFor this video, the text prompt is \"The lower extremities of humanoids consist of thighs, legs, and feet. The femur connects the hip to the knee, while the tibia and fibula form the lower leg, supporting the weight of the body and allowing various forms of locomotion, such as walking and running. Humanoids typically have feet designed for stability and balance, with arches that help absorb shock and distribute weight evenly during movement.\",\nall the frames of video are as follows:\n\n<image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> " }, { "from": "gpt", "value": "visual quality: 2\n temporal consistency: 1\n dynamic degree: 3\n text-to-video alignment: 1\n factual consistency: 1\n" } ]

[ "2000631_00.jpg", "2000631_01.jpg", "2000631_02.jpg", "2000631_03.jpg", "2000631_04.jpg", "2000631_05.jpg", "2000631_06.jpg", "2000631_07.jpg", "2000631_08.jpg", "2000631_09.jpg", "2000631_10.jpg", "2000631_11.jpg", "2000631_12.jpg", "2000631_13.jpg", "2000631_14.jpg", "2000631_15.jpg" ]

[ { "from": "human", "value": "\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos,\nplease watch the following frames of a given video and see the text prompt for generating the video,\nthen give scores from 5 different dimensions:\n(1) visual quality: the quality of the video in terms of clearness, resolution, brightness, and color\n(2) temporal consistency, both the consistency of objects or humans and the smoothness of motion or movements\n(3) dynamic degree, the degree of dynamic changes\n(4) text-to-video alignment, the alignment between the text prompt and the video content\n(5) factual consistency, the consistency of the video content with the common-sense and factual knowledge\n\nfor each dimension, output a float number from 1.0 to 4.0,\nthe higher the number is, the better the video performs in that sub-score, \nthe lowest 1.0 means Bad, the highest 4.0 means Perfect/Real (the video is like a real video)\nHere is an output example:\nvisual quality: 3.2\ntemporal consistency: 2.7\ndynamic degree: 4.0\ntext-to-video alignment: 2.3\nfactual consistency: 1.8\n\nFor this video, the text prompt is \"Muscular sweaty man running waving Turkish flag in his hand in rainy weather at night on the street\",\nall the frames of video are as follows:\n\n<image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> " }, { "from": "gpt", "value": "visual quality: 2\n temporal consistency: 2\n dynamic degree: 3\n text-to-video alignment: 2\n factual consistency: 2\n" } ]

[ "5000298_00.jpg", "5000298_01.jpg", "5000298_02.jpg", "5000298_03.jpg", "5000298_04.jpg", "5000298_05.jpg", "5000298_06.jpg", "5000298_07.jpg", "5000298_08.jpg", "5000298_09.jpg", "5000298_10.jpg", "5000298_11.jpg", "5000298_12.jpg", "5000298_13.jpg", "5000298_14.jpg", "5000298_15.jpg" ]

[ { "from": "human", "value": "\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos,\nplease watch the following frames of a given video and see the text prompt for generating the video,\nthen give scores from 5 different dimensions:\n(1) visual quality: the quality of the video in terms of clearness, resolution, brightness, and color\n(2) temporal consistency, both the consistency of objects or humans and the smoothness of motion or movements\n(3) dynamic degree, the degree of dynamic changes\n(4) text-to-video alignment, the alignment between the text prompt and the video content\n(5) factual consistency, the consistency of the video content with the common-sense and factual knowledge\n\nfor each dimension, output a float number from 1.0 to 4.0,\nthe higher the number is, the better the video performs in that sub-score, \nthe lowest 1.0 means Bad, the highest 4.0 means Perfect/Real (the video is like a real video)\nHere is an output example:\nvisual quality: 3.2\ntemporal consistency: 2.7\ndynamic degree: 4.0\ntext-to-video alignment: 2.3\nfactual consistency: 1.8\n\nFor this video, the text prompt is \"young woman smiling with her girlfriend\",\nall the frames of video are as follows:\n\n<image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> " }, { "from": "gpt", "value": "visual quality: 2\n temporal consistency: 1\n dynamic degree: 3\n text-to-video alignment: 3\n factual consistency: 1\n" } ]

[ "3002229_00.jpg", "3002229_01.jpg", "3002229_02.jpg", "3002229_03.jpg", "3002229_04.jpg", "3002229_05.jpg", "3002229_06.jpg", "3002229_07.jpg", "3002229_08.jpg", "3002229_09.jpg", "3002229_10.jpg", "3002229_11.jpg", "3002229_12.jpg", "3002229_13.jpg", "3002229_14.jpg", "3002229_15.jpg" ]

[ { "from": "human", "value": "\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos,\nplease watch the following frames of a given video and see the text prompt for generating the video,\nthen give scores from 5 different dimensions:\n(1) visual quality: the quality of the video in terms of clearness, resolution, brightness, and color\n(2) temporal consistency, both the consistency of objects or humans and the smoothness of motion or movements\n(3) dynamic degree, the degree of dynamic changes\n(4) text-to-video alignment, the alignment between the text prompt and the video content\n(5) factual consistency, the consistency of the video content with the common-sense and factual knowledge\n\nfor each dimension, output a float number from 1.0 to 4.0,\nthe higher the number is, the better the video performs in that sub-score, \nthe lowest 1.0 means Bad, the highest 4.0 means Perfect/Real (the video is like a real video)\nHere is an output example:\nvisual quality: 3.2\ntemporal consistency: 2.7\ndynamic degree: 4.0\ntext-to-video alignment: 2.3\nfactual consistency: 1.8\n\nFor this video, the text prompt is \"a sparrow flying high over the village\",\nall the frames of video are as follows:\n\n<image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> " }, { "from": "gpt", "value": "visual quality: 2\n temporal consistency: 1\n dynamic degree: 3\n text-to-video alignment: 1\n factual consistency: 1\n" } ]