Datasets:

---

hexuan21

/

VideoFeedback_r1k_bad8k

like 0

[ "0000475_00.jpg", "0000475_01.jpg", "0000475_02.jpg", "0000475_03.jpg", "0000475_04.jpg", "0000475_05.jpg", "0000475_06.jpg", "0000475_07.jpg", "0000475_08.jpg", "0000475_09.jpg", "0000475_10.jpg", "0000475_11.jpg", "0000475_12.jpg", "0000475_13.jpg", "0000475_14.jpg", "0000475_15.jpg", "0000475_16.jpg", "0000475_17.jpg", "0000475_18.jpg", "0000475_19.jpg", "0000475_20.jpg", "0000475_21.jpg", "0000475_22.jpg", "0000475_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 bkonde pretty woman early 50 s stood in the doorway of an enchanted cottage in the countryside, flowers grass swaying in the wind, birds, bees, rabbits, sun shining, clouds moving\",\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: 3\n text-to-video alignment: 0\n factual consistency: 3\n" } ]

[ "8002356_00.jpg", "8002356_01.jpg", "8002356_02.jpg", "8002356_03.jpg", "8002356_04.jpg", "8002356_05.jpg", "8002356_06.jpg", "8002356_07.jpg", "8002356_08.jpg", "8002356_09.jpg", "8002356_10.jpg", "8002356_11.jpg", "8002356_12.jpg", "8002356_13.jpg", "8002356_14.jpg", "8002356_15.jpg", "8002356_16.jpg", "8002356_17.jpg", "8002356_18.jpg", "8002356_19.jpg", "8002356_20.jpg", "8002356_21.jpg", "8002356_22.jpg", "8002356_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 black ford mustang drives fast on flat water\",\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: 0\n factual consistency: 1\n" } ]

[ "7004332_00.jpg", "7004332_01.jpg", "7004332_02.jpg", "7004332_03.jpg", "7004332_04.jpg", "7004332_05.jpg", "7004332_06.jpg", "7004332_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 \"nursery\",\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: 2\n dynamic degree: 3\n text-to-video alignment: 0\n factual consistency: 2\n" } ]

[ "3001462_00.jpg", "3001462_01.jpg", "3001462_02.jpg", "3001462_03.jpg", "3001462_04.jpg", "3001462_05.jpg", "3001462_06.jpg", "3001462_07.jpg", "3001462_08.jpg", "3001462_09.jpg", "3001462_10.jpg", "3001462_11.jpg", "3001462_12.jpg", "3001462_13.jpg", "3001462_14.jpg", "3001462_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 woman is singing at a microphone in a diner, in the style of olivier valsecchi, 32k uhd, marc simonetti, harlem renaissance, spectacular backdrops, , lively tavern scenes\",\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: 0\n factual consistency: 1\n" } ]

[ "1001005_00.jpg", "1001005_01.jpg", "1001005_02.jpg", "1001005_03.jpg", "1001005_04.jpg", "1001005_05.jpg", "1001005_06.jpg", "1001005_07.jpg", "1001005_08.jpg", "1001005_09.jpg", "1001005_10.jpg", "1001005_11.jpg", "1001005_12.jpg", "1001005_13.jpg", "1001005_14.jpg", "1001005_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 \"South Park Pajama Cartman dancing house\",\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: 0\n factual consistency: 1\n" } ]

[ "3006334_00.jpg", "3006334_01.jpg", "3006334_02.jpg", "3006334_03.jpg", "3006334_04.jpg", "3006334_05.jpg", "3006334_06.jpg", "3006334_07.jpg", "3006334_08.jpg", "3006334_09.jpg", "3006334_10.jpg", "3006334_11.jpg", "3006334_12.jpg", "3006334_13.jpg", "3006334_14.jpg", "3006334_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 \"romantic dinner in Rome where a couple is getting engaged. The scene is set in an elegant outdoor restaurant with a view of the Colosseum. The couple, a man and a woman, are sitting at a beautifully decorated table with candles and flowers. The man is presenting an engagement ring to the woman, and they both have expressions of love and surprise. The background shows the Colosseum illuminated at night, adding to the romantic ambiance. The scene is filled with warmth and happiness, capturing the moment of their engagement\",\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: 3\n dynamic degree: 3\n text-to-video alignment: 0\n factual consistency: 2\n" } ]

[ "1006551_00.jpg", "1006551_01.jpg", "1006551_02.jpg", "1006551_03.jpg", "1006551_04.jpg", "1006551_05.jpg", "1006551_06.jpg", "1006551_07.jpg", "1006551_08.jpg", "1006551_09.jpg", "1006551_10.jpg", "1006551_11.jpg", "1006551_12.jpg", "1006551_13.jpg", "1006551_14.jpg", "1006551_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 \"Draw cheerful kids with happy faces playing among meadow flowers and butterflies on a sunny summer day, golden ratio, fake detail, trending pixiv fanbox, acrylic palette knife, style of makoto shinkai studio ghibli genshin impact james gilleard greg rutkowski chiho aoshima, golden ratio, fake detail, trending pixiv fanbox, acrylic palette knife, style of makoto shinkai studio ghibli genshin impact james gilleard greg rutkowski chiho aoshima\",\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: 0\n factual consistency: 1\n" } ]

[ "6001632_00.jpg", "6001632_01.jpg", "6001632_02.jpg", "6001632_03.jpg", "6001632_04.jpg", "6001632_05.jpg", "6001632_06.jpg", "6001632_07.jpg", "6001632_08.jpg", "6001632_09.jpg", "6001632_10.jpg", "6001632_11.jpg", "6001632_12.jpg", "6001632_13.jpg", "6001632_14.jpg", "6001632_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 \"teens fashion waiting at train station\",\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: 0\n factual consistency: 1\n" } ]

[ "6001575_00.jpg", "6001575_01.jpg", "6001575_02.jpg", "6001575_03.jpg", "6001575_04.jpg", "6001575_05.jpg", "6001575_06.jpg", "6001575_07.jpg", "6001575_08.jpg", "6001575_09.jpg", "6001575_10.jpg", "6001575_11.jpg", "6001575_12.jpg", "6001575_13.jpg", "6001575_14.jpg", "6001575_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 \"Ludwig van beethoven listens to music on his headphones\",\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: 0\n factual consistency: 1\n" } ]

[ "0001542_00.jpg", "0001542_01.jpg", "0001542_02.jpg", "0001542_03.jpg", "0001542_04.jpg", "0001542_05.jpg", "0001542_06.jpg", "0001542_07.jpg", "0001542_08.jpg", "0001542_09.jpg", "0001542_10.jpg", "0001542_11.jpg", "0001542_12.jpg", "0001542_13.jpg", "0001542_14.jpg", "0001542_15.jpg", "0001542_16.jpg", "0001542_17.jpg", "0001542_18.jpg", "0001542_19.jpg", "0001542_20.jpg", "0001542_21.jpg", "0001542_22.jpg", "0001542_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 \"Show a moment of tension between the warriors three mighty warriors dressed in the traditional attire of Greece, Cyprus, and Turkey as they gaze upon the baklava trophy, highlighting their determined facial expressions while adorned in their authentic traditional garments.\",\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: 0\n factual consistency: 3\n" } ]

[ "2000213_00.jpg", "2000213_01.jpg", "2000213_02.jpg", "2000213_03.jpg", "2000213_04.jpg", "2000213_05.jpg", "2000213_06.jpg", "2000213_07.jpg", "2000213_08.jpg", "2000213_09.jpg", "2000213_10.jpg", "2000213_11.jpg", "2000213_12.jpg", "2000213_13.jpg", "2000213_14.jpg", "2000213_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 \"colorful melting art hyper detail award winning\",\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: 0\n factual consistency: 1\n" } ]

[ "3000761_00.jpg", "3000761_01.jpg", "3000761_02.jpg", "3000761_03.jpg", "3000761_04.jpg", "3000761_05.jpg", "3000761_06.jpg", "3000761_07.jpg", "3000761_08.jpg", "3000761_09.jpg", "3000761_10.jpg", "3000761_11.jpg", "3000761_12.jpg", "3000761_13.jpg", "3000761_14.jpg", "3000761_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 gorgeously rendered papercraft world of a coral reef, rife with colorful fish and sea creatures.\",\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: 0\n factual consistency: 2\n" } ]

[ "6000172_00.jpg", "6000172_01.jpg", "6000172_02.jpg", "6000172_03.jpg", "6000172_04.jpg", "6000172_05.jpg", "6000172_06.jpg", "6000172_07.jpg", "6000172_08.jpg", "6000172_09.jpg", "6000172_10.jpg", "6000172_11.jpg", "6000172_12.jpg", "6000172_13.jpg", "6000172_14.jpg", "6000172_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 logo with I,Y,Y letters by using black and blue\",\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: 3\n dynamic degree: 3\n text-to-video alignment: 0\n factual consistency: 1\n" } ]

[ "6001508_00.jpg", "6001508_01.jpg", "6001508_02.jpg", "6001508_03.jpg", "6001508_04.jpg", "6001508_05.jpg", "6001508_06.jpg", "6001508_07.jpg", "6001508_08.jpg", "6001508_09.jpg", "6001508_10.jpg", "6001508_11.jpg", "6001508_12.jpg", "6001508_13.jpg", "6001508_14.jpg", "6001508_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 \"an ia robot walking in a park\",\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: 0\n factual consistency: 1\n" } ]

[ "0005465_00.jpg", "0005465_01.jpg", "0005465_02.jpg", "0005465_03.jpg", "0005465_04.jpg", "0005465_05.jpg", "0005465_06.jpg", "0005465_07.jpg", "0005465_08.jpg", "0005465_09.jpg", "0005465_10.jpg", "0005465_11.jpg", "0005465_12.jpg", "0005465_13.jpg", "0005465_14.jpg", "0005465_15.jpg", "0005465_16.jpg", "0005465_17.jpg", "0005465_18.jpg", "0005465_19.jpg", "0005465_20.jpg", "0005465_21.jpg", "0005465_22.jpg", "0005465_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 \"person walking in the dark city at night in the foreground with an imaginary friend floating on his side of a reddish color with orange, lights focused on the road\",\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: 0\n factual consistency: 3\n" } ]

[ "3005863_00.jpg", "3005863_01.jpg", "3005863_02.jpg", "3005863_03.jpg", "3005863_04.jpg", "3005863_05.jpg", "3005863_06.jpg", "3005863_07.jpg", "3005863_08.jpg", "3005863_09.jpg", "3005863_10.jpg", "3005863_11.jpg", "3005863_12.jpg", "3005863_13.jpg", "3005863_14.jpg", "3005863_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 dog interviewing a cat that stole its acoustic guitar\",\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: 0\n factual consistency: 1\n" } ]

[ "8000379_00.jpg", "8000379_01.jpg", "8000379_02.jpg", "8000379_03.jpg", "8000379_04.jpg", "8000379_05.jpg", "8000379_06.jpg", "8000379_07.jpg", "8000379_08.jpg", "8000379_09.jpg", "8000379_10.jpg", "8000379_11.jpg", "8000379_12.jpg", "8000379_13.jpg", "8000379_14.jpg", "8000379_15.jpg", "8000379_16.jpg", "8000379_17.jpg", "8000379_18.jpg", "8000379_19.jpg", "8000379_20.jpg", "8000379_21.jpg", "8000379_22.jpg", "8000379_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 \"studio ghibli movie, hand drawn animation, aerial view of a seaside Japanese town with traditional houses\",\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: 0\n factual consistency: 1\n" } ]

[ "8002721_00.jpg", "8002721_01.jpg", "8002721_02.jpg", "8002721_03.jpg", "8002721_04.jpg", "8002721_05.jpg", "8002721_06.jpg", "8002721_07.jpg", "8002721_08.jpg", "8002721_09.jpg", "8002721_10.jpg", "8002721_11.jpg", "8002721_12.jpg", "8002721_13.jpg", "8002721_14.jpg", "8002721_15.jpg", "8002721_16.jpg", "8002721_17.jpg", "8002721_18.jpg", "8002721_19.jpg", "8002721_20.jpg", "8002721_21.jpg", "8002721_22.jpg", "8002721_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 boy plays video games in the year 1455\",\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: 0\n factual consistency: 1\n" } ]

[ "1002732_00.jpg", "1002732_01.jpg", "1002732_02.jpg", "1002732_03.jpg", "1002732_04.jpg", "1002732_05.jpg", "1002732_06.jpg", "1002732_07.jpg", "1002732_08.jpg", "1002732_09.jpg", "1002732_10.jpg", "1002732_11.jpg", "1002732_12.jpg", "1002732_13.jpg", "1002732_14.jpg", "1002732_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 \"Jason Statham fights a Werewolf in a dark forest\",\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: 0\n factual consistency: 1\n" } ]

[ "3000870_00.jpg", "3000870_01.jpg", "3000870_02.jpg", "3000870_03.jpg", "3000870_04.jpg", "3000870_05.jpg", "3000870_06.jpg", "3000870_07.jpg", "3000870_08.jpg", "3000870_09.jpg", "3000870_10.jpg", "3000870_11.jpg", "3000870_12.jpg", "3000870_13.jpg", "3000870_14.jpg", "3000870_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 \"vr people marching in a 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> " }, { "from": "gpt", "value": "visual quality: 2\n temporal consistency: 1\n dynamic degree: 2\n text-to-video alignment: 0\n factual consistency: 1\n" } ]

[ "3005036_00.jpg", "3005036_01.jpg", "3005036_02.jpg", "3005036_03.jpg", "3005036_04.jpg", "3005036_05.jpg", "3005036_06.jpg", "3005036_07.jpg", "3005036_08.jpg", "3005036_09.jpg", "3005036_10.jpg", "3005036_11.jpg", "3005036_12.jpg", "3005036_13.jpg", "3005036_14.jpg", "3005036_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 futuristic medical protontherapy particle accelerator room depicted in a Pixar 3D animation 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: 2\n text-to-video alignment: 0\n factual consistency: 1\n" } ]

[ "7000827_00.jpg", "7000827_01.jpg", "7000827_02.jpg", "7000827_03.jpg", "7000827_04.jpg", "7000827_05.jpg", "7000827_06.jpg", "7000827_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 \"beautiful child looking at the crescent moon\",\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: 0\n factual consistency: 2\n" } ]

[ "8000618_00.jpg", "8000618_01.jpg", "8000618_02.jpg", "8000618_03.jpg", "8000618_04.jpg", "8000618_05.jpg", "8000618_06.jpg", "8000618_07.jpg", "8000618_08.jpg", "8000618_09.jpg", "8000618_10.jpg", "8000618_11.jpg", "8000618_12.jpg", "8000618_13.jpg", "8000618_14.jpg", "8000618_15.jpg", "8000618_16.jpg", "8000618_17.jpg", "8000618_18.jpg", "8000618_19.jpg", "8000618_20.jpg", "8000618_21.jpg", "8000618_22.jpg", "8000618_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 sky that is made up of the inside of a lava lamp, the horizon is on fire, angels are flying in the sky, a man floats on a tube in the water drinking a margarita while balancing a boom box on his belly. aliens come and abduct people in the background. cities are frozen in the distance.\",\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: 0\n factual consistency: 1\n" } ]

[ "1001199_00.jpg", "1001199_01.jpg", "1001199_02.jpg", "1001199_03.jpg", "1001199_04.jpg", "1001199_05.jpg", "1001199_06.jpg", "1001199_07.jpg", "1001199_08.jpg", "1001199_09.jpg", "1001199_10.jpg", "1001199_11.jpg", "1001199_12.jpg", "1001199_13.jpg", "1001199_14.jpg", "1001199_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 man driving a sport cart in the streets of El Salvador\",\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: 0\n factual consistency: 1\n" } ]

[ "5000675_00.jpg", "5000675_01.jpg", "5000675_02.jpg", "5000675_03.jpg", "5000675_04.jpg", "5000675_05.jpg", "5000675_06.jpg", "5000675_07.jpg", "5000675_08.jpg", "5000675_09.jpg", "5000675_10.jpg", "5000675_11.jpg", "5000675_12.jpg", "5000675_13.jpg", "5000675_14.jpg", "5000675_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 \"Spiderman watching over new York from top of empire State Building in night with moon and stars up in the sky. HD\",\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: 0\n factual consistency: 3\n" } ]

[ "1004289_00.jpg", "1004289_01.jpg", "1004289_02.jpg", "1004289_03.jpg", "1004289_04.jpg", "1004289_05.jpg", "1004289_06.jpg", "1004289_07.jpg", "1004289_08.jpg", "1004289_09.jpg", "1004289_10.jpg", "1004289_11.jpg", "1004289_12.jpg", "1004289_13.jpg", "1004289_14.jpg", "1004289_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 Link from Zelda holding a blaster, the blaster shoots a green bolt, his green cape and hair blows in the wind, in the background flames billow\",\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: 0\n factual consistency: 1\n" } ]

[ "1001021_00.jpg", "1001021_01.jpg", "1001021_02.jpg", "1001021_03.jpg", "1001021_04.jpg", "1001021_05.jpg", "1001021_06.jpg", "1001021_07.jpg", "1001021_08.jpg", "1001021_09.jpg", "1001021_10.jpg", "1001021_11.jpg", "1001021_12.jpg", "1001021_13.jpg", "1001021_14.jpg", "1001021_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 boy dancing with a Pepsi bottle\",\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: 0\n factual consistency: 1\n" } ]

[ "1000302_00.jpg", "1000302_01.jpg", "1000302_02.jpg", "1000302_03.jpg", "1000302_04.jpg", "1000302_05.jpg", "1000302_06.jpg", "1000302_07.jpg", "1000302_08.jpg", "1000302_09.jpg", "1000302_10.jpg", "1000302_11.jpg", "1000302_12.jpg", "1000302_13.jpg", "1000302_14.jpg", "1000302_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 \"Participants straining to hear whispers in the dark. Horror scary 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: 1\n temporal consistency: 2\n dynamic degree: 3\n text-to-video alignment: 0\n factual consistency: 1\n" } ]

[ "1004766_00.jpg", "1004766_01.jpg", "1004766_02.jpg", "1004766_03.jpg", "1004766_04.jpg", "1004766_05.jpg", "1004766_06.jpg", "1004766_07.jpg", "1004766_08.jpg", "1004766_09.jpg", "1004766_10.jpg", "1004766_11.jpg", "1004766_12.jpg", "1004766_13.jpg", "1004766_14.jpg", "1004766_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 \"As the girls climbed the stairs, they heard a strange noise, a terrifying whisper. They turned around, but there was no one there. Their flashlights began to flicker, casting eerie shadows on the walls.\",\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: 0\n factual consistency: 2\n" } ]

[ "1000008_00.jpg", "1000008_01.jpg", "1000008_02.jpg", "1000008_03.jpg", "1000008_04.jpg", "1000008_05.jpg", "1000008_06.jpg", "1000008_07.jpg", "1000008_08.jpg", "1000008_09.jpg", "1000008_10.jpg", "1000008_11.jpg", "1000008_12.jpg", "1000008_13.jpg", "1000008_14.jpg", "1000008_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 \"\",\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: 0\n factual consistency: 1\n" } ]

[ "1000565_00.jpg", "1000565_01.jpg", "1000565_02.jpg", "1000565_03.jpg", "1000565_04.jpg", "1000565_05.jpg", "1000565_06.jpg", "1000565_07.jpg", "1000565_08.jpg", "1000565_09.jpg", "1000565_10.jpg", "1000565_11.jpg", "1000565_12.jpg", "1000565_13.jpg", "1000565_14.jpg", "1000565_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 young woman walks towards the educational building, the building is visible in front, trees grow along the edges of the path, the day is sunny\",\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: 0\n factual consistency: 1\n" } ]

[ "0001678_00.jpg", "0001678_01.jpg", "0001678_02.jpg", "0001678_03.jpg", "0001678_04.jpg", "0001678_05.jpg", "0001678_06.jpg", "0001678_07.jpg", "0001678_08.jpg", "0001678_09.jpg", "0001678_10.jpg", "0001678_11.jpg", "0001678_12.jpg", "0001678_13.jpg", "0001678_14.jpg", "0001678_15.jpg", "0001678_16.jpg", "0001678_17.jpg", "0001678_18.jpg", "0001678_19.jpg", "0001678_20.jpg", "0001678_21.jpg", "0001678_22.jpg", "0001678_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 \"looking through a window, camera zooming in on birds at a feeder\",\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: 0\n factual consistency: 1\n" } ]

[ "6001131_00.jpg", "6001131_01.jpg", "6001131_02.jpg", "6001131_03.jpg", "6001131_04.jpg", "6001131_05.jpg", "6001131_06.jpg", "6001131_07.jpg", "6001131_08.jpg", "6001131_09.jpg", "6001131_10.jpg", "6001131_11.jpg", "6001131_12.jpg", "6001131_13.jpg", "6001131_14.jpg", "6001131_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 \"crying with really sad emotions slowly zoming in\",\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: 0\n factual consistency: 1\n" } ]

[ "0001150_00.jpg", "0001150_01.jpg", "0001150_02.jpg", "0001150_03.jpg", "0001150_04.jpg", "0001150_05.jpg", "0001150_06.jpg", "0001150_07.jpg", "0001150_08.jpg", "0001150_09.jpg", "0001150_10.jpg", "0001150_11.jpg", "0001150_12.jpg", "0001150_13.jpg", "0001150_14.jpg", "0001150_15.jpg", "0001150_16.jpg", "0001150_17.jpg", "0001150_18.jpg", "0001150_19.jpg", "0001150_20.jpg", "0001150_21.jpg", "0001150_22.jpg", "0001150_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 video that two couples drinking coffee in coffee shop in a resort in mountain\",\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: 3\n text-to-video alignment: 0\n factual consistency: 3\n" } ]

[ "4002936_00.jpg", "4002936_01.jpg", "4002936_02.jpg", "4002936_03.jpg", "4002936_04.jpg", "4002936_05.jpg", "4002936_06.jpg", "4002936_07.jpg", "4002936_08.jpg", "4002936_09.jpg", "4002936_10.jpg", "4002936_11.jpg", "4002936_12.jpg", "4002936_13.jpg", "4002936_14.jpg", "4002936_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 \"When the soldiers looked up, they saw that a crow was slowly flying away carrying the necklace.\",\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: 0\n factual consistency: 1\n" } ]

[ "0002166_00.jpg", "0002166_01.jpg", "0002166_02.jpg", "0002166_03.jpg", "0002166_04.jpg", "0002166_05.jpg", "0002166_06.jpg", "0002166_07.jpg", "0002166_08.jpg", "0002166_09.jpg", "0002166_10.jpg", "0002166_11.jpg", "0002166_12.jpg", "0002166_13.jpg", "0002166_14.jpg", "0002166_15.jpg", "0002166_16.jpg", "0002166_17.jpg", "0002166_18.jpg", "0002166_19.jpg", "0002166_20.jpg", "0002166_21.jpg", "0002166_22.jpg", "0002166_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 \"Boy with glasses drinking frappuccino on cafe terrace at night\",\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: 0\n factual consistency: 3\n" } ]

[ "7001694_00.jpg", "7001694_01.jpg", "7001694_02.jpg", "7001694_03.jpg", "7001694_04.jpg", "7001694_05.jpg", "7001694_06.jpg", "7001694_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 \"charater Hatsune Miku in anime Evangelion in anime style .There is furniture around, and they are chilling on their couch while looking at their phone. Character is also swinging their leg casually.\",\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: 2\n dynamic degree: 3\n text-to-video alignment: 0\n factual consistency: 3\n" } ]

[ "1000129_00.jpg", "1000129_01.jpg", "1000129_02.jpg", "1000129_03.jpg", "1000129_04.jpg", "1000129_05.jpg", "1000129_06.jpg", "1000129_07.jpg", "1000129_08.jpg", "1000129_09.jpg", "1000129_10.jpg", "1000129_11.jpg", "1000129_12.jpg", "1000129_13.jpg", "1000129_14.jpg", "1000129_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 \"there is a little girl walking in the streets in the morning while chewing gum while she is watching some architecture 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: 1\n temporal consistency: 2\n dynamic degree: 3\n text-to-video alignment: 0\n factual consistency: 1\n" } ]

[ "8000311_00.jpg", "8000311_01.jpg", "8000311_02.jpg", "8000311_03.jpg", "8000311_04.jpg", "8000311_05.jpg", "8000311_06.jpg", "8000311_07.jpg", "8000311_08.jpg", "8000311_09.jpg", "8000311_10.jpg", "8000311_11.jpg", "8000311_12.jpg", "8000311_13.jpg", "8000311_14.jpg", "8000311_15.jpg", "8000311_16.jpg", "8000311_17.jpg", "8000311_18.jpg", "8000311_19.jpg", "8000311_20.jpg", "8000311_21.jpg", "8000311_22.jpg", "8000311_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 \"video of a cow eating ice cream\",\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: 0\n factual consistency: 1\n" } ]

[ "1002105_00.jpg", "1002105_01.jpg", "1002105_02.jpg", "1002105_03.jpg", "1002105_04.jpg", "1002105_05.jpg", "1002105_06.jpg", "1002105_07.jpg", "1002105_08.jpg", "1002105_09.jpg", "1002105_10.jpg", "1002105_11.jpg", "1002105_12.jpg", "1002105_13.jpg", "1002105_14.jpg", "1002105_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 vase on the right of scissors, front 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: 1\n temporal consistency: 1\n dynamic degree: 3\n text-to-video alignment: 0\n factual consistency: 1\n" } ]

[ "8003680_00.jpg", "8003680_01.jpg", "8003680_02.jpg", "8003680_03.jpg", "8003680_04.jpg", "8003680_05.jpg", "8003680_06.jpg", "8003680_07.jpg", "8003680_08.jpg", "8003680_09.jpg", "8003680_10.jpg", "8003680_11.jpg", "8003680_12.jpg", "8003680_13.jpg", "8003680_14.jpg", "8003680_15.jpg", "8003680_16.jpg", "8003680_17.jpg", "8003680_18.jpg", "8003680_19.jpg", "8003680_20.jpg", "8003680_21.jpg", "8003680_22.jpg", "8003680_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 smiling buda sweating holding a eletric fan in one hands and a umbrella in the other hand with a huge sun behind him in one half and a huge storm on the other half of the background and some birds flying in the background, highly detailed\",\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: 0\n factual consistency: 1\n" } ]

[ "3005457_00.jpg", "3005457_01.jpg", "3005457_02.jpg", "3005457_03.jpg", "3005457_04.jpg", "3005457_05.jpg", "3005457_06.jpg", "3005457_07.jpg", "3005457_08.jpg", "3005457_09.jpg", "3005457_10.jpg", "3005457_11.jpg", "3005457_12.jpg", "3005457_13.jpg", "3005457_14.jpg", "3005457_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 \"There was a child named Samir, who lived in a quiet neighborhood with his parents and siblings\",\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: 0\n factual consistency: 3\n" } ]

[ "4000497_00.jpg", "4000497_01.jpg", "4000497_02.jpg", "4000497_03.jpg", "4000497_04.jpg", "4000497_05.jpg", "4000497_06.jpg", "4000497_07.jpg", "4000497_08.jpg", "4000497_09.jpg", "4000497_10.jpg", "4000497_11.jpg", "4000497_12.jpg", "4000497_13.jpg", "4000497_14.jpg", "4000497_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 \"trees, flowers, fauna and flora moves by the breezes, realistic, motion , 8k,9:16\",\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: 0\n factual consistency: 3\n" } ]

[ "6002868_00.jpg", "6002868_01.jpg", "6002868_02.jpg", "6002868_03.jpg", "6002868_04.jpg", "6002868_05.jpg", "6002868_06.jpg", "6002868_07.jpg", "6002868_08.jpg", "6002868_09.jpg", "6002868_10.jpg", "6002868_11.jpg", "6002868_12.jpg", "6002868_13.jpg", "6002868_14.jpg", "6002868_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 \"Spiderman works as a PIZZA delivery boy\",\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: 3\n dynamic degree: 3\n text-to-video alignment: 0\n factual consistency: 1\n" } ]

[ "4000772_00.jpg", "4000772_01.jpg", "4000772_02.jpg", "4000772_03.jpg", "4000772_04.jpg", "4000772_05.jpg", "4000772_06.jpg", "4000772_07.jpg", "4000772_08.jpg", "4000772_09.jpg", "4000772_10.jpg", "4000772_11.jpg", "4000772_12.jpg", "4000772_13.jpg", "4000772_14.jpg", "4000772_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 scary hooded figure walks in a dark room.\",\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: 0\n factual consistency: 3\n" } ]

[ "3002386_00.jpg", "3002386_01.jpg", "3002386_02.jpg", "3002386_03.jpg", "3002386_04.jpg", "3002386_05.jpg", "3002386_06.jpg", "3002386_07.jpg", "3002386_08.jpg", "3002386_09.jpg", "3002386_10.jpg", "3002386_11.jpg", "3002386_12.jpg", "3002386_13.jpg", "3002386_14.jpg", "3002386_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 man riding amotor bike on fire\",\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: 0\n factual consistency: 2\n" } ]

[ "7001693_00.jpg", "7001693_01.jpg", "7001693_02.jpg", "7001693_03.jpg", "7001693_04.jpg", "7001693_05.jpg", "7001693_06.jpg", "7001693_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 \"1950’s super 8 footage of creature split in two\",\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: 0\n factual consistency: 3\n" } ]

[ "5001909_00.jpg", "5001909_01.jpg", "5001909_02.jpg", "5001909_03.jpg", "5001909_04.jpg", "5001909_05.jpg", "5001909_06.jpg", "5001909_07.jpg", "5001909_08.jpg", "5001909_09.jpg", "5001909_10.jpg", "5001909_11.jpg", "5001909_12.jpg", "5001909_13.jpg", "5001909_14.jpg", "5001909_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 salesperson writes a check to a customer in a grocery store\",\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: 0\n factual consistency: 3\n" } ]

[ "4000010_00.jpg", "4000010_01.jpg", "4000010_02.jpg", "4000010_03.jpg", "4000010_04.jpg", "4000010_05.jpg", "4000010_06.jpg", "4000010_07.jpg", "4000010_08.jpg", "4000010_09.jpg", "4000010_10.jpg", "4000010_11.jpg", "4000010_12.jpg", "4000010_13.jpg", "4000010_14.jpg", "4000010_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 music video telling a story in the style of metallica music videos\",\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: 0\n factual consistency: 1\n" } ]

[ "3001090_00.jpg", "3001090_01.jpg", "3001090_02.jpg", "3001090_03.jpg", "3001090_04.jpg", "3001090_05.jpg", "3001090_06.jpg", "3001090_07.jpg", "3001090_08.jpg", "3001090_09.jpg", "3001090_10.jpg", "3001090_11.jpg", "3001090_12.jpg", "3001090_13.jpg", "3001090_14.jpg", "3001090_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 an image that symbolizes the process of amplifying controversial content on social media.\",\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: 0\n factual consistency: 1\n" } ]

[ "0004365_00.jpg", "0004365_01.jpg", "0004365_02.jpg", "0004365_03.jpg", "0004365_04.jpg", "0004365_05.jpg", "0004365_06.jpg", "0004365_07.jpg", "0004365_08.jpg", "0004365_09.jpg", "0004365_10.jpg", "0004365_11.jpg", "0004365_12.jpg", "0004365_13.jpg", "0004365_14.jpg", "0004365_15.jpg", "0004365_16.jpg", "0004365_17.jpg", "0004365_18.jpg", "0004365_19.jpg", "0004365_20.jpg", "0004365_21.jpg", "0004365_22.jpg", "0004365_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 cat eating a human, cat power\",\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: 0\n factual consistency: 2\n" } ]

[ "2002911_00.jpg", "2002911_01.jpg", "2002911_02.jpg", "2002911_03.jpg", "2002911_04.jpg", "2002911_05.jpg", "2002911_06.jpg", "2002911_07.jpg", "2002911_08.jpg", "2002911_09.jpg", "2002911_10.jpg", "2002911_11.jpg", "2002911_12.jpg", "2002911_13.jpg", "2002911_14.jpg", "2002911_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 \"an agent in a classic suit descends from a helicopter on a rope, classic suit, helicopter, rope\",\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: 0\n factual consistency: 3\n" } ]

[ "8002362_00.jpg", "8002362_01.jpg", "8002362_02.jpg", "8002362_03.jpg", "8002362_04.jpg", "8002362_05.jpg", "8002362_06.jpg", "8002362_07.jpg", "8002362_08.jpg", "8002362_09.jpg", "8002362_10.jpg", "8002362_11.jpg", "8002362_12.jpg", "8002362_13.jpg", "8002362_14.jpg", "8002362_15.jpg", "8002362_16.jpg", "8002362_17.jpg", "8002362_18.jpg", "8002362_19.jpg", "8002362_20.jpg", "8002362_21.jpg", "8002362_22.jpg", "8002362_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 \"run through the streets of new york\",\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: 2\n dynamic degree: 2\n text-to-video alignment: 0\n factual consistency: 1\n" } ]

[ "1006791_00.jpg", "1006791_01.jpg", "1006791_02.jpg", "1006791_03.jpg", "1006791_04.jpg", "1006791_05.jpg", "1006791_06.jpg", "1006791_07.jpg", "1006791_08.jpg", "1006791_09.jpg", "1006791_10.jpg", "1006791_11.jpg", "1006791_12.jpg", "1006791_13.jpg", "1006791_14.jpg", "1006791_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 \"burn effect silent film person with octopuspot struggles,high resolution,noise basquit color palette\",\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: 0\n factual consistency: 2\n" } ]

[ "0000191_00.jpg", "0000191_01.jpg", "0000191_02.jpg", "0000191_03.jpg", "0000191_04.jpg", "0000191_05.jpg", "0000191_06.jpg", "0000191_07.jpg", "0000191_08.jpg", "0000191_09.jpg", "0000191_10.jpg", "0000191_11.jpg", "0000191_12.jpg", "0000191_13.jpg", "0000191_14.jpg", "0000191_15.jpg", "0000191_16.jpg", "0000191_17.jpg", "0000191_18.jpg", "0000191_19.jpg", "0000191_20.jpg", "0000191_21.jpg", "0000191_22.jpg", "0000191_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 \"Male college student, without glasses, working as a programmer to make money, comic style, the computer is a laptop and the system is Fedora\",\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: 0\n factual consistency: 3\n" } ]

[ "0001654_00.jpg", "0001654_01.jpg", "0001654_02.jpg", "0001654_03.jpg", "0001654_04.jpg", "0001654_05.jpg", "0001654_06.jpg", "0001654_07.jpg", "0001654_08.jpg", "0001654_09.jpg", "0001654_10.jpg", "0001654_11.jpg", "0001654_12.jpg", "0001654_13.jpg", "0001654_14.jpg", "0001654_15.jpg", "0001654_16.jpg", "0001654_17.jpg", "0001654_18.jpg", "0001654_19.jpg", "0001654_20.jpg", "0001654_21.jpg", "0001654_22.jpg", "0001654_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 \"squares moving inside into a void\",\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: 2\n text-to-video alignment: 0\n factual consistency: 3\n" } ]

[ "6001853_00.jpg", "6001853_01.jpg", "6001853_02.jpg", "6001853_03.jpg", "6001853_04.jpg", "6001853_05.jpg", "6001853_06.jpg", "6001853_07.jpg", "6001853_08.jpg", "6001853_09.jpg", "6001853_10.jpg", "6001853_11.jpg", "6001853_12.jpg", "6001853_13.jpg", "6001853_14.jpg", "6001853_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 \"mechanical human heart beating at 1 beat per second.\",\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: 0\n factual consistency: 1\n" } ]

[ "1000393_00.jpg", "1000393_01.jpg", "1000393_02.jpg", "1000393_03.jpg", "1000393_04.jpg", "1000393_05.jpg", "1000393_06.jpg", "1000393_07.jpg", "1000393_08.jpg", "1000393_09.jpg", "1000393_10.jpg", "1000393_11.jpg", "1000393_12.jpg", "1000393_13.jpg", "1000393_14.jpg", "1000393_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 market in a fantastic village, movie film, fantasy, hight quality, hight detailed, 4K, movie, film, cinematic movie\",\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: 0\n factual consistency: 1\n" } ]

[ "2002450_00.jpg", "2002450_01.jpg", "2002450_02.jpg", "2002450_03.jpg", "2002450_04.jpg", "2002450_05.jpg", "2002450_06.jpg", "2002450_07.jpg", "2002450_08.jpg", "2002450_09.jpg", "2002450_10.jpg", "2002450_11.jpg", "2002450_12.jpg", "2002450_13.jpg", "2002450_14.jpg", "2002450_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 boy standing on rocks, during a storm, 20 seconds\",\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: 0\n factual consistency: 3\n" } ]

[ "1006527_00.jpg", "1006527_01.jpg", "1006527_02.jpg", "1006527_03.jpg", "1006527_04.jpg", "1006527_05.jpg", "1006527_06.jpg", "1006527_07.jpg", "1006527_08.jpg", "1006527_09.jpg", "1006527_10.jpg", "1006527_11.jpg", "1006527_12.jpg", "1006527_13.jpg", "1006527_14.jpg", "1006527_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 woman hesitated for a moment, then accepted his offer. They sat down at a corner table, their eyes meeting for the first time. Her name was Meera, and she had been caught in the rain while waiting for a friend who had stood her up.\",\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: 0\n factual consistency: 1\n" } ]

[ "7004401_00.jpg", "7004401_01.jpg", "7004401_02.jpg", "7004401_03.jpg", "7004401_04.jpg", "7004401_05.jpg", "7004401_06.jpg", "7004401_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 \"Four camera operators are running away\",\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: 0\n factual consistency: 1\n" } ]

[ "6001569_00.jpg", "6001569_01.jpg", "6001569_02.jpg", "6001569_03.jpg", "6001569_04.jpg", "6001569_05.jpg", "6001569_06.jpg", "6001569_07.jpg", "6001569_08.jpg", "6001569_09.jpg", "6001569_10.jpg", "6001569_11.jpg", "6001569_12.jpg", "6001569_13.jpg", "6001569_14.jpg", "6001569_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 \"speed up 100 000 years history in 3 sec shot, nature landscape metamorphosing,\",\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: 0\n factual consistency: 1\n" } ]

[ "7000520_00.jpg", "7000520_01.jpg", "7000520_02.jpg", "7000520_03.jpg", "7000520_04.jpg", "7000520_05.jpg", "7000520_06.jpg", "7000520_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 \"only backgeound the smoke will blow , no character will not move include head hands and more\",\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: 0\n factual consistency: 3\n" } ]

[ "6000476_00.jpg", "6000476_01.jpg", "6000476_02.jpg", "6000476_03.jpg", "6000476_04.jpg", "6000476_05.jpg", "6000476_06.jpg", "6000476_07.jpg", "6000476_08.jpg", "6000476_09.jpg", "6000476_10.jpg", "6000476_11.jpg", "6000476_12.jpg", "6000476_13.jpg", "6000476_14.jpg", "6000476_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 \"chill animation, young man with mask, walking\",\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: 0\n factual consistency: 1\n" } ]

[ "1001107_00.jpg", "1001107_01.jpg", "1001107_02.jpg", "1001107_03.jpg", "1001107_04.jpg", "1001107_05.jpg", "1001107_06.jpg", "1001107_07.jpg", "1001107_08.jpg", "1001107_09.jpg", "1001107_10.jpg", "1001107_11.jpg", "1001107_12.jpg", "1001107_13.jpg", "1001107_14.jpg", "1001107_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 \"following sci fi warrior in armour coming through the forest at sunset with mountains in background with high intensity and quick motion\",\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: 0\n factual consistency: 1\n" } ]

[ "4004788_00.jpg", "4004788_01.jpg", "4004788_02.jpg", "4004788_03.jpg", "4004788_04.jpg", "4004788_05.jpg", "4004788_06.jpg", "4004788_07.jpg", "4004788_08.jpg", "4004788_09.jpg", "4004788_10.jpg", "4004788_11.jpg", "4004788_12.jpg", "4004788_13.jpg", "4004788_14.jpg", "4004788_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 \"Twinkle, a curious star, gazing down at Earth.\",\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: 0\n factual consistency: 3\n" } ]

[ "1000247_00.jpg", "1000247_01.jpg", "1000247_02.jpg", "1000247_03.jpg", "1000247_04.jpg", "1000247_05.jpg", "1000247_06.jpg", "1000247_07.jpg", "1000247_08.jpg", "1000247_09.jpg", "1000247_10.jpg", "1000247_11.jpg", "1000247_12.jpg", "1000247_13.jpg", "1000247_14.jpg", "1000247_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 cute happy Corgi playing in park, sunset\",\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: 0\n factual consistency: 1\n" } ]

[ "0003786_00.jpg", "0003786_01.jpg", "0003786_02.jpg", "0003786_03.jpg", "0003786_04.jpg", "0003786_05.jpg", "0003786_06.jpg", "0003786_07.jpg", "0003786_08.jpg", "0003786_09.jpg", "0003786_10.jpg", "0003786_11.jpg", "0003786_12.jpg", "0003786_13.jpg", "0003786_14.jpg", "0003786_15.jpg", "0003786_16.jpg", "0003786_17.jpg", "0003786_18.jpg", "0003786_19.jpg", "0003786_20.jpg", "0003786_21.jpg", "0003786_22.jpg", "0003786_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 \"fast paced grime music video set in a gritty housing estate in London in 1999, rainy, architecture, dark grimey street, dynamic lighting, cinematic, ultra realistic, highly detailed, apartment\",\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: 3\n text-to-video alignment: 0\n factual consistency: 1\n" } ]

[ "8003668_00.jpg", "8003668_01.jpg", "8003668_02.jpg", "8003668_03.jpg", "8003668_04.jpg", "8003668_05.jpg", "8003668_06.jpg", "8003668_07.jpg", "8003668_08.jpg", "8003668_09.jpg", "8003668_10.jpg", "8003668_11.jpg", "8003668_12.jpg", "8003668_13.jpg", "8003668_14.jpg", "8003668_15.jpg", "8003668_16.jpg", "8003668_17.jpg", "8003668_18.jpg", "8003668_19.jpg", "8003668_20.jpg", "8003668_21.jpg", "8003668_22.jpg", "8003668_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 young boy lay in his bed at night,4k,animated,night effect\",\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: 2\n dynamic degree: 3\n text-to-video alignment: 0\n factual consistency: 1\n" } ]

[ "0006717_00.jpg", "0006717_01.jpg", "0006717_02.jpg", "0006717_03.jpg", "0006717_04.jpg", "0006717_05.jpg", "0006717_06.jpg", "0006717_07.jpg", "0006717_08.jpg", "0006717_09.jpg", "0006717_10.jpg", "0006717_11.jpg", "0006717_12.jpg", "0006717_13.jpg", "0006717_14.jpg", "0006717_15.jpg", "0006717_16.jpg", "0006717_17.jpg", "0006717_18.jpg", "0006717_19.jpg", "0006717_20.jpg", "0006717_21.jpg", "0006717_22.jpg", "0006717_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 Zen Temple atop a tall mountain\",\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: 3\n text-to-video alignment: 0\n factual consistency: 2\n" } ]

[ "7004673_00.jpg", "7004673_01.jpg", "7004673_02.jpg", "7004673_03.jpg", "7004673_04.jpg", "7004673_05.jpg", "7004673_06.jpg", "7004673_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 massive riot in front of a industrial complex\",\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: 0\n factual consistency: 3\n" } ]

[ "2005777_00.jpg", "2005777_01.jpg", "2005777_02.jpg", "2005777_03.jpg", "2005777_04.jpg", "2005777_05.jpg", "2005777_06.jpg", "2005777_07.jpg", "2005777_08.jpg", "2005777_09.jpg", "2005777_10.jpg", "2005777_11.jpg", "2005777_12.jpg", "2005777_13.jpg", "2005777_14.jpg", "2005777_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 \"\",\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: 0\n factual consistency: 2\n" } ]

[ "0000866_00.jpg", "0000866_01.jpg", "0000866_02.jpg", "0000866_03.jpg", "0000866_04.jpg", "0000866_05.jpg", "0000866_06.jpg", "0000866_07.jpg", "0000866_08.jpg", "0000866_09.jpg", "0000866_10.jpg", "0000866_11.jpg", "0000866_12.jpg", "0000866_13.jpg", "0000866_14.jpg", "0000866_15.jpg", "0000866_16.jpg", "0000866_17.jpg", "0000866_18.jpg", "0000866_19.jpg", "0000866_20.jpg", "0000866_21.jpg", "0000866_22.jpg", "0000866_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 alien blending in naturally with new york city, paranoia thriller style, 35mm film\",\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: 0\n factual consistency: 3\n" } ]

[ "2003591_00.jpg", "2003591_01.jpg", "2003591_02.jpg", "2003591_03.jpg", "2003591_04.jpg", "2003591_05.jpg", "2003591_06.jpg", "2003591_07.jpg", "2003591_08.jpg", "2003591_09.jpg", "2003591_10.jpg", "2003591_11.jpg", "2003591_12.jpg", "2003591_13.jpg", "2003591_14.jpg", "2003591_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 \"Team Spirit lose Riyadh Masters 2023, tournament, dota 2, arena\",\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: 0\n factual consistency: 3\n" } ]

[ "0001365_00.jpg", "0001365_01.jpg", "0001365_02.jpg", "0001365_03.jpg", "0001365_04.jpg", "0001365_05.jpg", "0001365_06.jpg", "0001365_07.jpg", "0001365_08.jpg", "0001365_09.jpg", "0001365_10.jpg", "0001365_11.jpg", "0001365_12.jpg", "0001365_13.jpg", "0001365_14.jpg", "0001365_15.jpg", "0001365_16.jpg", "0001365_17.jpg", "0001365_18.jpg", "0001365_19.jpg", "0001365_20.jpg", "0001365_21.jpg", "0001365_22.jpg", "0001365_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 cute fairy girl alone in the jungle.\",\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: 0\n factual consistency: 3\n" } ]

[ "3003224_00.jpg", "3003224_01.jpg", "3003224_02.jpg", "3003224_03.jpg", "3003224_04.jpg", "3003224_05.jpg", "3003224_06.jpg", "3003224_07.jpg", "3003224_08.jpg", "3003224_09.jpg", "3003224_10.jpg", "3003224_11.jpg", "3003224_12.jpg", "3003224_13.jpg", "3003224_14.jpg", "3003224_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 \"drumset hardware falling from the sky, blue sky mountains\",\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: 0\n factual consistency: 1\n" } ]

[ "1001175_00.jpg", "1001175_01.jpg", "1001175_02.jpg", "1001175_03.jpg", "1001175_04.jpg", "1001175_05.jpg", "1001175_06.jpg", "1001175_07.jpg", "1001175_08.jpg", "1001175_09.jpg", "1001175_10.jpg", "1001175_11.jpg", "1001175_12.jpg", "1001175_13.jpg", "1001175_14.jpg", "1001175_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 \"Iron Man and Batman are fighting, with Iron Man firing a laser attack.\",\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: 0\n factual consistency: 1\n" } ]

[ "4003468_00.jpg", "4003468_01.jpg", "4003468_02.jpg", "4003468_03.jpg", "4003468_04.jpg", "4003468_05.jpg", "4003468_06.jpg", "4003468_07.jpg", "4003468_08.jpg", "4003468_09.jpg", "4003468_10.jpg", "4003468_11.jpg", "4003468_12.jpg", "4003468_13.jpg", "4003468_14.jpg", "4003468_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 \"add pink lighting coming off bottle\",\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: 0\n factual consistency: 1\n" } ]

[ "6002221_00.jpg", "6002221_01.jpg", "6002221_02.jpg", "6002221_03.jpg", "6002221_04.jpg", "6002221_05.jpg", "6002221_06.jpg", "6002221_07.jpg", "6002221_08.jpg", "6002221_09.jpg", "6002221_10.jpg", "6002221_11.jpg", "6002221_12.jpg", "6002221_13.jpg", "6002221_14.jpg", "6002221_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 \"horror movie style black and white clouds and eagles\",\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: 0\n factual consistency: 1\n" } ]

[ "6003391_00.jpg", "6003391_01.jpg", "6003391_02.jpg", "6003391_03.jpg", "6003391_04.jpg", "6003391_05.jpg", "6003391_06.jpg", "6003391_07.jpg", "6003391_08.jpg", "6003391_09.jpg", "6003391_10.jpg", "6003391_11.jpg", "6003391_12.jpg", "6003391_13.jpg", "6003391_14.jpg", "6003391_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 \"Meditating man on snow during the day, realistic art, 9:16\",\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: 0\n factual consistency: 1\n" } ]

[ "7000685_00.jpg", "7000685_01.jpg", "7000685_02.jpg", "7000685_03.jpg", "7000685_04.jpg", "7000685_05.jpg", "7000685_06.jpg", "7000685_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 \"cat talking to audiance with smile\",\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: 0\n factual consistency: 3\n" } ]

[ "3001730_00.jpg", "3001730_01.jpg", "3001730_02.jpg", "3001730_03.jpg", "3001730_04.jpg", "3001730_05.jpg", "3001730_06.jpg", "3001730_07.jpg", "3001730_08.jpg", "3001730_09.jpg", "3001730_10.jpg", "3001730_11.jpg", "3001730_12.jpg", "3001730_13.jpg", "3001730_14.jpg", "3001730_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 \"Pharaoh and his army realizing that they are doomed, depicted with a look of panic and fear on their faces.\",\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: 0\n factual consistency: 2\n" } ]

[ "1004500_00.jpg", "1004500_01.jpg", "1004500_02.jpg", "1004500_03.jpg", "1004500_04.jpg", "1004500_05.jpg", "1004500_06.jpg", "1004500_07.jpg", "1004500_08.jpg", "1004500_09.jpg", "1004500_10.jpg", "1004500_11.jpg", "1004500_12.jpg", "1004500_13.jpg", "1004500_14.jpg", "1004500_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 group of people fly to the ocean of dreams, their faces filled with happy smiles\",\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: 0\n factual consistency: 1\n" } ]

[ "s000553_00.jpg", "s000553_01.jpg", "s000553_02.jpg", "s000553_03.jpg", "s000553_04.jpg", "s000553_05.jpg", "s000553_06.jpg", "s000553_07.jpg", "s000553_08.jpg", "s000553_09.jpg", "s000553_10.jpg", "s000553_11.jpg", "s000553_12.jpg", "s000553_13.jpg", "s000553_14.jpg", "s000553_15.jpg", "s000553_16.jpg", "s000553_17.jpg", "s000553_18.jpg", "s000553_19.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 \"emji is bought at vending machine.\",\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> " }, { "from": "gpt", "value": "visual quality: 4\n temporal consistency: 4\n dynamic degree: 4\n text-to-video alignment: 0\n factual consistency: 4\n" } ]

[ "8002507_00.jpg", "8002507_01.jpg", "8002507_02.jpg", "8002507_03.jpg", "8002507_04.jpg", "8002507_05.jpg", "8002507_06.jpg", "8002507_07.jpg", "8002507_08.jpg", "8002507_09.jpg", "8002507_10.jpg", "8002507_11.jpg", "8002507_12.jpg", "8002507_13.jpg", "8002507_14.jpg", "8002507_15.jpg", "8002507_16.jpg", "8002507_17.jpg", "8002507_18.jpg", "8002507_19.jpg", "8002507_20.jpg", "8002507_21.jpg", "8002507_22.jpg", "8002507_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 \"Girl partying in a spaceship with lots of aliens, anime 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> <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: 0\n factual consistency: 1\n" } ]

[ "0000523_00.jpg", "0000523_01.jpg", "0000523_02.jpg", "0000523_03.jpg", "0000523_04.jpg", "0000523_05.jpg", "0000523_06.jpg", "0000523_07.jpg", "0000523_08.jpg", "0000523_09.jpg", "0000523_10.jpg", "0000523_11.jpg", "0000523_12.jpg", "0000523_13.jpg", "0000523_14.jpg", "0000523_15.jpg", "0000523_16.jpg", "0000523_17.jpg", "0000523_18.jpg", "0000523_19.jpg", "0000523_20.jpg", "0000523_21.jpg", "0000523_22.jpg", "0000523_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 boy walking in forest Whitaker his father\",\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: 0\n factual consistency: 3\n" } ]

[ "3001160_00.jpg", "3001160_01.jpg", "3001160_02.jpg", "3001160_03.jpg", "3001160_04.jpg", "3001160_05.jpg", "3001160_06.jpg", "3001160_07.jpg", "3001160_08.jpg", "3001160_09.jpg", "3001160_10.jpg", "3001160_11.jpg", "3001160_12.jpg", "3001160_13.jpg", "3001160_14.jpg", "3001160_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 \"steam train, in the mountains, night, smoke,\",\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: 0\n factual consistency: 3\n" } ]

[ "3002125_00.jpg", "3002125_01.jpg", "3002125_02.jpg", "3002125_03.jpg", "3002125_04.jpg", "3002125_05.jpg", "3002125_06.jpg", "3002125_07.jpg", "3002125_08.jpg", "3002125_09.jpg", "3002125_10.jpg", "3002125_11.jpg", "3002125_12.jpg", "3002125_13.jpg", "3002125_14.jpg", "3002125_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 polar bear is playing guitar\",\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: 1\n text-to-video alignment: 0\n factual consistency: 1\n" } ]

[ "0005578_00.jpg", "0005578_01.jpg", "0005578_02.jpg", "0005578_03.jpg", "0005578_04.jpg", "0005578_05.jpg", "0005578_06.jpg", "0005578_07.jpg", "0005578_08.jpg", "0005578_09.jpg", "0005578_10.jpg", "0005578_11.jpg", "0005578_12.jpg", "0005578_13.jpg", "0005578_14.jpg", "0005578_15.jpg", "0005578_16.jpg", "0005578_17.jpg", "0005578_18.jpg", "0005578_19.jpg", "0005578_20.jpg", "0005578_21.jpg", "0005578_22.jpg", "0005578_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 UAE man vector art looking worried and thinking\",\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: 0\n factual consistency: 2\n" } ]

[ "0002193_00.jpg", "0002193_01.jpg", "0002193_02.jpg", "0002193_03.jpg", "0002193_04.jpg", "0002193_05.jpg", "0002193_06.jpg", "0002193_07.jpg", "0002193_08.jpg", "0002193_09.jpg", "0002193_10.jpg", "0002193_11.jpg", "0002193_12.jpg", "0002193_13.jpg", "0002193_14.jpg", "0002193_15.jpg", "0002193_16.jpg", "0002193_17.jpg", "0002193_18.jpg", "0002193_19.jpg", "0002193_20.jpg", "0002193_21.jpg", "0002193_22.jpg", "0002193_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 Indian boy driving a car\",\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: 0\n factual consistency: 3\n" } ]

[ "0006664_00.jpg", "0006664_01.jpg", "0006664_02.jpg", "0006664_03.jpg", "0006664_04.jpg", "0006664_05.jpg", "0006664_06.jpg", "0006664_07.jpg", "0006664_08.jpg", "0006664_09.jpg", "0006664_10.jpg", "0006664_11.jpg", "0006664_12.jpg", "0006664_13.jpg", "0006664_14.jpg", "0006664_15.jpg", "0006664_16.jpg", "0006664_17.jpg", "0006664_18.jpg", "0006664_19.jpg", "0006664_20.jpg", "0006664_21.jpg", "0006664_22.jpg", "0006664_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 black and white cat is sleeping in a bed\",\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: 0\n factual consistency: 3\n" } ]

[ "4000447_00.jpg", "4000447_01.jpg", "4000447_02.jpg", "4000447_03.jpg", "4000447_04.jpg", "4000447_05.jpg", "4000447_06.jpg", "4000447_07.jpg", "4000447_08.jpg", "4000447_09.jpg", "4000447_10.jpg", "4000447_11.jpg", "4000447_12.jpg", "4000447_13.jpg", "4000447_14.jpg", "4000447_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 \"Driving corpses in Xiangxi, a man wearing Taoist clothing is driving a group of entities walking\",\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: 0\n factual consistency: 3\n" } ]

[ "5000978_00.jpg", "5000978_01.jpg", "5000978_02.jpg", "5000978_03.jpg", "5000978_04.jpg", "5000978_05.jpg", "5000978_06.jpg", "5000978_07.jpg", "5000978_08.jpg", "5000978_09.jpg", "5000978_10.jpg", "5000978_11.jpg", "5000978_12.jpg", "5000978_13.jpg", "5000978_14.jpg", "5000978_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 of a mob of people in front of a burning vehicle at a riot Message: DIER DEOL (Font: MODERN)\",\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: 0\n factual consistency: 2\n" } ]

[ "2003755_00.jpg", "2003755_01.jpg", "2003755_02.jpg", "2003755_03.jpg", "2003755_04.jpg", "2003755_05.jpg", "2003755_06.jpg", "2003755_07.jpg", "2003755_08.jpg", "2003755_09.jpg", "2003755_10.jpg", "2003755_11.jpg", "2003755_12.jpg", "2003755_13.jpg", "2003755_14.jpg", "2003755_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 snow pile turned into a real Pikachu\",\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: 0\n factual consistency: 2\n" } ]

[ "2005264_00.jpg", "2005264_01.jpg", "2005264_02.jpg", "2005264_03.jpg", "2005264_04.jpg", "2005264_05.jpg", "2005264_06.jpg", "2005264_07.jpg", "2005264_08.jpg", "2005264_09.jpg", "2005264_10.jpg", "2005264_11.jpg", "2005264_12.jpg", "2005264_13.jpg", "2005264_14.jpg", "2005264_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 man sitting to beach far way the light house\",\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: 0\n factual consistency: 1\n" } ]

[ "3006930_00.jpg", "3006930_01.jpg", "3006930_02.jpg", "3006930_03.jpg", "3006930_04.jpg", "3006930_05.jpg", "3006930_06.jpg", "3006930_07.jpg", "3006930_08.jpg", "3006930_09.jpg", "3006930_10.jpg", "3006930_11.jpg", "3006930_12.jpg", "3006930_13.jpg", "3006930_14.jpg", "3006930_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 \"ninja blade swift speed through 3 people slow motion hyper focused hyper detailed\",\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: 0\n factual consistency: 1\n" } ]

[ "5000589_00.jpg", "5000589_01.jpg", "5000589_02.jpg", "5000589_03.jpg", "5000589_04.jpg", "5000589_05.jpg", "5000589_06.jpg", "5000589_07.jpg", "5000589_08.jpg", "5000589_09.jpg", "5000589_10.jpg", "5000589_11.jpg", "5000589_12.jpg", "5000589_13.jpg", "5000589_14.jpg", "5000589_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 \"Show an animated sequence of their friendship evolving into love through changing facial expressions, from playful teasing to moments of shared intimacy.\",\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: 2\n text-to-video alignment: 0\n factual consistency: 2\n" } ]

[ "5000522_00.jpg", "5000522_01.jpg", "5000522_02.jpg", "5000522_03.jpg", "5000522_04.jpg", "5000522_05.jpg", "5000522_06.jpg", "5000522_07.jpg", "5000522_08.jpg", "5000522_09.jpg", "5000522_10.jpg", "5000522_11.jpg", "5000522_12.jpg", "5000522_13.jpg", "5000522_14.jpg", "5000522_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 \"Lily attempting to climb a tall tree, capturing the thrill of the moment.Animated cartoon\",\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: 2\n text-to-video alignment: 0\n factual consistency: 1\n" } ]

[ "2001958_00.jpg", "2001958_01.jpg", "2001958_02.jpg", "2001958_03.jpg", "2001958_04.jpg", "2001958_05.jpg", "2001958_06.jpg", "2001958_07.jpg", "2001958_08.jpg", "2001958_09.jpg", "2001958_10.jpg", "2001958_11.jpg", "2001958_12.jpg", "2001958_13.jpg", "2001958_14.jpg", "2001958_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 golden and shiny 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: 3\n dynamic degree: 2\n text-to-video alignment: 0\n factual consistency: 3\n" } ]

[ "7003673_00.jpg", "7003673_01.jpg", "7003673_02.jpg", "7003673_03.jpg", "7003673_04.jpg", "7003673_05.jpg", "7003673_06.jpg", "7003673_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 \"spanish football team Message: spain (Font: MODERN)\",\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: 2\n text-to-video alignment: 0\n factual consistency: 1\n" } ]