Datasets:

hexuan21

/

VideoFeedback_bad8k

like 0

[ "5000546_00.jpg", "5000546_01.jpg", "5000546_02.jpg", "5000546_03.jpg", "5000546_04.jpg", "5000546_05.jpg", "5000546_06.jpg", "5000546_07.jpg", "5000546_08.jpg", "5000546_09.jpg", "5000546_10.jpg", "5000546_11.jpg", "5000546_12.jpg", "5000546_13.jpg", "5000546_14.jpg", "5000546_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 \"penguin playing drums Message: CHILI LINDO (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: 3\n temporal consistency: 2\n dynamic degree: 3\n text-to-video alignment: 1\n factual consistency: 2\n" } ]

[ "4000203_00.jpg", "4000203_01.jpg", "4000203_02.jpg", "4000203_03.jpg", "4000203_04.jpg", "4000203_05.jpg", "4000203_06.jpg", "4000203_07.jpg", "4000203_08.jpg", "4000203_09.jpg", "4000203_10.jpg", "4000203_11.jpg", "4000203_12.jpg", "4000203_13.jpg", "4000203_14.jpg", "4000203_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 laughing with a young child\",\nall the frames of video are as follows:\n\n<image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> " }, { "from": "gpt", "value": "visual quality: 3\n temporal consistency: 2\n dynamic degree: 3\n text-to-video alignment: 3\n factual consistency: 3\n" } ]

[ "2001829_00.jpg", "2001829_01.jpg", "2001829_02.jpg", "2001829_03.jpg", "2001829_04.jpg", "2001829_05.jpg", "2001829_06.jpg", "2001829_07.jpg", "2001829_08.jpg", "2001829_09.jpg", "2001829_10.jpg", "2001829_11.jpg", "2001829_12.jpg", "2001829_13.jpg", "2001829_14.jpg", "2001829_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 moon cake with eyes and 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: 3\n temporal consistency: 3\n dynamic degree: 2\n text-to-video alignment: 3\n factual consistency: 3\n" } ]

[ "0001172_00.jpg", "0001172_01.jpg", "0001172_02.jpg", "0001172_03.jpg", "0001172_04.jpg", "0001172_05.jpg", "0001172_06.jpg", "0001172_07.jpg", "0001172_08.jpg", "0001172_09.jpg", "0001172_10.jpg", "0001172_11.jpg", "0001172_12.jpg", "0001172_13.jpg", "0001172_14.jpg", "0001172_15.jpg", "0001172_16.jpg", "0001172_17.jpg", "0001172_18.jpg", "0001172_19.jpg", "0001172_20.jpg", "0001172_21.jpg", "0001172_22.jpg", "0001172_23.jpg" ]

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

[ "4000676_00.jpg", "4000676_01.jpg", "4000676_02.jpg", "4000676_03.jpg", "4000676_04.jpg", "4000676_05.jpg", "4000676_06.jpg", "4000676_07.jpg", "4000676_08.jpg", "4000676_09.jpg", "4000676_10.jpg", "4000676_11.jpg", "4000676_12.jpg", "4000676_13.jpg", "4000676_14.jpg", "4000676_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 vibrant scene of a skateboarder performing an incredible trick in a halfpipe surrounded by blooming trees. Spring colors fill the environment, with flower petals swirling in the air as the skateboarder conquers the ramp.\",\nall the frames of video are as follows:\n\n<image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> " }, { "from": "gpt", "value": "visual quality: 2\n temporal consistency: 2\n dynamic degree: 3\n text-to-video alignment: 1\n factual consistency: 3\n" } ]

[ "0002010_00.jpg", "0002010_01.jpg", "0002010_02.jpg", "0002010_03.jpg", "0002010_04.jpg", "0002010_05.jpg", "0002010_06.jpg", "0002010_07.jpg", "0002010_08.jpg", "0002010_09.jpg", "0002010_10.jpg", "0002010_11.jpg", "0002010_12.jpg", "0002010_13.jpg", "0002010_14.jpg", "0002010_15.jpg", "0002010_16.jpg", "0002010_17.jpg", "0002010_18.jpg", "0002010_19.jpg", "0002010_20.jpg", "0002010_21.jpg", "0002010_22.jpg", "0002010_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 \"gold coin falling and spinning in slow 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> <image> <image> <image> <image> <image> <image> <image> <image> " }, { "from": "gpt", "value": "visual quality: 3\n temporal consistency: 3\n dynamic degree: 3\n text-to-video alignment: 3\n factual consistency: 3\n" } ]

[ "0002077_00.jpg", "0002077_01.jpg", "0002077_02.jpg", "0002077_03.jpg", "0002077_04.jpg", "0002077_05.jpg", "0002077_06.jpg", "0002077_07.jpg", "0002077_08.jpg", "0002077_09.jpg", "0002077_10.jpg", "0002077_11.jpg", "0002077_12.jpg", "0002077_13.jpg", "0002077_14.jpg", "0002077_15.jpg", "0002077_16.jpg", "0002077_17.jpg", "0002077_18.jpg", "0002077_19.jpg", "0002077_20.jpg", "0002077_21.jpg", "0002077_22.jpg", "0002077_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 futuristic samurai in the robotic suite in cyberpunk style is standing exausted after a fight with 20 soldiers who are now dead and laying on the ground all around him. He is holding a katana sword in one hand while the edge of katana is resting on the ground\",\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: 2\n factual consistency: 3\n" } ]

[ "2000134_00.jpg", "2000134_01.jpg", "2000134_02.jpg", "2000134_03.jpg", "2000134_04.jpg", "2000134_05.jpg", "2000134_06.jpg", "2000134_07.jpg", "2000134_08.jpg", "2000134_09.jpg", "2000134_10.jpg", "2000134_11.jpg", "2000134_12.jpg", "2000134_13.jpg", "2000134_14.jpg", "2000134_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 pink teddy bear, move his ears and blink an eye\",\nall the frames of video are as follows:\n\n<image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> " }, { "from": "gpt", "value": "visual quality: 3\n temporal consistency: 3\n dynamic degree: 2\n text-to-video alignment: 3\n factual consistency: 3\n" } ]

[ "6001560_00.jpg", "6001560_01.jpg", "6001560_02.jpg", "6001560_03.jpg", "6001560_04.jpg", "6001560_05.jpg", "6001560_06.jpg", "6001560_07.jpg", "6001560_08.jpg", "6001560_09.jpg", "6001560_10.jpg", "6001560_11.jpg", "6001560_12.jpg", "6001560_13.jpg", "6001560_14.jpg", "6001560_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 \"China Kung Fu Panda flies in the air.\",\nall the frames of video are as follows:\n\n<image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> " }, { "from": "gpt", "value": "visual quality: 1\n temporal consistency: 1\n dynamic degree: 3\n text-to-video alignment: 2\n factual consistency: 1\n" } ]

[ "1000563_00.jpg", "1000563_01.jpg", "1000563_02.jpg", "1000563_03.jpg", "1000563_04.jpg", "1000563_05.jpg", "1000563_06.jpg", "1000563_07.jpg", "1000563_08.jpg", "1000563_09.jpg", "1000563_10.jpg", "1000563_11.jpg", "1000563_12.jpg", "1000563_13.jpg", "1000563_14.jpg", "1000563_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 \"picture illustration for the book: medieval princesses and poetesses\",\nall the frames of video are as follows:\n\n<image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> " }, { "from": "gpt", "value": "visual quality: 1\n temporal consistency: 1\n dynamic degree: 3\n text-to-video alignment: 2\n factual consistency: 1\n" } ]

[ "4001181_00.jpg", "4001181_01.jpg", "4001181_02.jpg", "4001181_03.jpg", "4001181_04.jpg", "4001181_05.jpg", "4001181_06.jpg", "4001181_07.jpg", "4001181_08.jpg", "4001181_09.jpg", "4001181_10.jpg", "4001181_11.jpg", "4001181_12.jpg", "4001181_13.jpg", "4001181_14.jpg", "4001181_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 happy beautiful princess in a white gown holding hand of a happy handsome prince wearing white Royal attire standing in a castle balcony\",\nall the frames of video are as follows:\n\n<image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> " }, { "from": "gpt", "value": "visual quality: 3\n temporal consistency: 3\n dynamic degree: 3\n text-to-video alignment: 3\n factual consistency: 3\n" } ]

[ "0001058_00.jpg", "0001058_01.jpg", "0001058_02.jpg", "0001058_03.jpg", "0001058_04.jpg", "0001058_05.jpg", "0001058_06.jpg", "0001058_07.jpg", "0001058_08.jpg", "0001058_09.jpg", "0001058_10.jpg", "0001058_11.jpg", "0001058_12.jpg", "0001058_13.jpg", "0001058_14.jpg", "0001058_15.jpg", "0001058_16.jpg", "0001058_17.jpg", "0001058_18.jpg", "0001058_19.jpg", "0001058_20.jpg", "0001058_21.jpg", "0001058_22.jpg", "0001058_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 young girl study on her table in her room near the window and study usiing her notebook in the night, the weather is rainy and the place is calm and her white cat sleeps on the 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: 3\n temporal consistency: 3\n dynamic degree: 2\n text-to-video alignment: 3\n factual consistency: 3\n" } ]

[ "2001178_00.jpg", "2001178_01.jpg", "2001178_02.jpg", "2001178_03.jpg", "2001178_04.jpg", "2001178_05.jpg", "2001178_06.jpg", "2001178_07.jpg", "2001178_08.jpg", "2001178_09.jpg", "2001178_10.jpg", "2001178_11.jpg", "2001178_12.jpg", "2001178_13.jpg", "2001178_14.jpg", "2001178_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 \"Guardian of the light fighting the evil in a heroic fantasy 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: 3\n temporal consistency: 3\n dynamic degree: 3\n text-to-video alignment: 2\n factual consistency: 3\n" } ]

[ "2001853_00.jpg", "2001853_01.jpg", "2001853_02.jpg", "2001853_03.jpg", "2001853_04.jpg", "2001853_05.jpg", "2001853_06.jpg", "2001853_07.jpg", "2001853_08.jpg", "2001853_09.jpg", "2001853_10.jpg", "2001853_11.jpg", "2001853_12.jpg", "2001853_13.jpg", "2001853_14.jpg", "2001853_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 \"shadows moving on staircase 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> " }, { "from": "gpt", "value": "visual quality: 3\n temporal consistency: 3\n dynamic degree: 3\n text-to-video alignment: 3\n factual consistency: 3\n" } ]

[ "6000195_00.jpg", "6000195_01.jpg", "6000195_02.jpg", "6000195_03.jpg", "6000195_04.jpg", "6000195_05.jpg", "6000195_06.jpg", "6000195_07.jpg", "6000195_08.jpg", "6000195_09.jpg", "6000195_10.jpg", "6000195_11.jpg", "6000195_12.jpg", "6000195_13.jpg", "6000195_14.jpg", "6000195_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 mechanical sphere floating over a pink desert\",\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: 3\n factual consistency: 1\n" } ]

[ "7001566_00.jpg", "7001566_01.jpg", "7001566_02.jpg", "7001566_03.jpg", "7001566_04.jpg", "7001566_05.jpg", "7001566_06.jpg", "7001566_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 \"Animal growl in the center, text do not touch\",\nall the frames of video are as follows:\n\n<image> <image> <image> <image> <image> <image> <image> <image> " }, { "from": "gpt", "value": "visual quality: 2\n temporal consistency: 2\n dynamic degree: 3\n text-to-video alignment: 3\n factual consistency: 3\n" } ]

[ "1001144_00.jpg", "1001144_01.jpg", "1001144_02.jpg", "1001144_03.jpg", "1001144_04.jpg", "1001144_05.jpg", "1001144_06.jpg", "1001144_07.jpg", "1001144_08.jpg", "1001144_09.jpg", "1001144_10.jpg", "1001144_11.jpg", "1001144_12.jpg", "1001144_13.jpg", "1001144_14.jpg", "1001144_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 \"udyr league of lagends cartoon animation playing in a pc\",\nall the frames of video are as follows:\n\n<image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> " }, { "from": "gpt", "value": "visual quality: 1\n temporal consistency: 1\n dynamic degree: 3\n text-to-video alignment: 2\n factual consistency: 1\n" } ]

[ "4000573_00.jpg", "4000573_01.jpg", "4000573_02.jpg", "4000573_03.jpg", "4000573_04.jpg", "4000573_05.jpg", "4000573_06.jpg", "4000573_07.jpg", "4000573_08.jpg", "4000573_09.jpg", "4000573_10.jpg", "4000573_11.jpg", "4000573_12.jpg", "4000573_13.jpg", "4000573_14.jpg", "4000573_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 \"Yuki encounters a powerful sorcerer testing his skills. Yuki dazzles everyone with his remarkable magical prowess, proving his strength , 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> " }, { "from": "gpt", "value": "visual quality: 3\n temporal consistency: 2\n dynamic degree: 3\n text-to-video alignment: 3\n factual consistency: 1\n" } ]

[ "3002922_00.jpg", "3002922_01.jpg", "3002922_02.jpg", "3002922_03.jpg", "3002922_04.jpg", "3002922_05.jpg", "3002922_06.jpg", "3002922_07.jpg", "3002922_08.jpg", "3002922_09.jpg", "3002922_10.jpg", "3002922_11.jpg", "3002922_12.jpg", "3002922_13.jpg", "3002922_14.jpg", "3002922_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 team is discussing work in the office. There are computers and whiteboards, and the atmosphere is lively. , Anime Key Visual\",\nall the frames of video are as follows:\n\n<image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> " }, { "from": "gpt", "value": "visual quality: 1\n temporal consistency: 2\n dynamic degree: 3\n text-to-video alignment: 2\n factual consistency: 2\n" } ]

[ "0001385_00.jpg", "0001385_01.jpg", "0001385_02.jpg", "0001385_03.jpg", "0001385_04.jpg", "0001385_05.jpg", "0001385_06.jpg", "0001385_07.jpg", "0001385_08.jpg", "0001385_09.jpg", "0001385_10.jpg", "0001385_11.jpg", "0001385_12.jpg", "0001385_13.jpg", "0001385_14.jpg", "0001385_15.jpg", "0001385_16.jpg", "0001385_17.jpg", "0001385_18.jpg", "0001385_19.jpg", "0001385_20.jpg", "0001385_21.jpg", "0001385_22.jpg", "0001385_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 \"vintage scene with house at distance Message: Purush Mandal (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> <image> <image> <image> <image> <image> <image> <image> <image> " }, { "from": "gpt", "value": "visual quality: 3\n temporal consistency: 3\n dynamic degree: 2\n text-to-video alignment: 3\n factual consistency: 3\n" } ]

[ "6000679_00.jpg", "6000679_01.jpg", "6000679_02.jpg", "6000679_03.jpg", "6000679_04.jpg", "6000679_05.jpg", "6000679_06.jpg", "6000679_07.jpg", "6000679_08.jpg", "6000679_09.jpg", "6000679_10.jpg", "6000679_11.jpg", "6000679_12.jpg", "6000679_13.jpg", "6000679_14.jpg", "6000679_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 \"person attacked by a vampire in the towers of serrano, in valencia,8k,hyperrealistic\",\nall the frames of video are as follows:\n\n<image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> " }, { "from": "gpt", "value": "visual quality: 2\n temporal consistency: 2\n dynamic degree: 3\n text-to-video alignment: 3\n factual consistency: 1\n" } ]

[ "4001620_00.jpg", "4001620_01.jpg", "4001620_02.jpg", "4001620_03.jpg", "4001620_04.jpg", "4001620_05.jpg", "4001620_06.jpg", "4001620_07.jpg", "4001620_08.jpg", "4001620_09.jpg", "4001620_10.jpg", "4001620_11.jpg", "4001620_12.jpg", "4001620_13.jpg", "4001620_14.jpg", "4001620_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 \"Ukrainian warrior flying and protect Ukraine from the russian rockets, high quality, high speed\",\nall the frames of video are as follows:\n\n<image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> " }, { "from": "gpt", "value": "visual quality: 2\n temporal consistency: 2\n dynamic degree: 3\n text-to-video alignment: 3\n factual consistency: 3\n" } ]

[ "2002076_00.jpg", "2002076_01.jpg", "2002076_02.jpg", "2002076_03.jpg", "2002076_04.jpg", "2002076_05.jpg", "2002076_06.jpg", "2002076_07.jpg", "2002076_08.jpg", "2002076_09.jpg", "2002076_10.jpg", "2002076_11.jpg", "2002076_12.jpg", "2002076_13.jpg", "2002076_14.jpg", "2002076_15.jpg" ]

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

[ "8000120_00.jpg", "8000120_01.jpg", "8000120_02.jpg", "8000120_03.jpg", "8000120_04.jpg", "8000120_05.jpg", "8000120_06.jpg", "8000120_07.jpg", "8000120_08.jpg", "8000120_09.jpg", "8000120_10.jpg", "8000120_11.jpg", "8000120_12.jpg", "8000120_13.jpg", "8000120_14.jpg", "8000120_15.jpg", "8000120_16.jpg", "8000120_17.jpg", "8000120_18.jpg", "8000120_19.jpg", "8000120_20.jpg", "8000120_21.jpg", "8000120_22.jpg", "8000120_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 \"electro man feeding the fox on the table\",\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: 1\n factual consistency: 1\n" } ]

[ "1001187_00.jpg", "1001187_01.jpg", "1001187_02.jpg", "1001187_03.jpg", "1001187_04.jpg", "1001187_05.jpg", "1001187_06.jpg", "1001187_07.jpg", "1001187_08.jpg", "1001187_09.jpg", "1001187_10.jpg", "1001187_11.jpg", "1001187_12.jpg", "1001187_13.jpg", "1001187_14.jpg", "1001187_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 caveman finds a red button in the cave. Close up a d very detailed . 4K HD. Cinematic style.\",\nall the frames of video are as follows:\n\n<image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> " }, { "from": "gpt", "value": "visual quality: 1\n temporal consistency: 1\n dynamic degree: 3\n text-to-video alignment: 2\n factual consistency: 1\n" } ]

[ "4001655_00.jpg", "4001655_01.jpg", "4001655_02.jpg", "4001655_03.jpg", "4001655_04.jpg", "4001655_05.jpg", "4001655_06.jpg", "4001655_07.jpg", "4001655_08.jpg", "4001655_09.jpg", "4001655_10.jpg", "4001655_11.jpg", "4001655_12.jpg", "4001655_13.jpg", "4001655_14.jpg", "4001655_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 \"realistic little girl sitting alone in collapsed building\",\nall the frames of video are as follows:\n\n<image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> " }, { "from": "gpt", "value": "visual quality: 2\n temporal consistency: 3\n dynamic degree: 3\n text-to-video alignment: 3\n factual consistency: 3\n" } ]

[ "7000516_00.jpg", "7000516_01.jpg", "7000516_02.jpg", "7000516_03.jpg", "7000516_04.jpg", "7000516_05.jpg", "7000516_06.jpg", "7000516_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 \"Brazilian woman holding baby on her lap, northeast, backlands, heat landscape\",\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: 3\n factual consistency: 3\n" } ]

[ "1002291_00.jpg", "1002291_01.jpg", "1002291_02.jpg", "1002291_03.jpg", "1002291_04.jpg", "1002291_05.jpg", "1002291_06.jpg", "1002291_07.jpg", "1002291_08.jpg", "1002291_09.jpg", "1002291_10.jpg", "1002291_11.jpg", "1002291_12.jpg", "1002291_13.jpg", "1002291_14.jpg", "1002291_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 cinematic shot of a group of soldiers trapped on the beach in the style of Avatar movie, surrounded by enemy forces, desperation and the chaos of war, IMAX camera, super detailing\",\nall the frames of video are as follows:\n\n<image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> " }, { "from": "gpt", "value": "visual quality: 1\n temporal consistency: 1\n dynamic degree: 3\n text-to-video alignment: 2\n factual consistency: 1\n" } ]

[ "2001601_00.jpg", "2001601_01.jpg", "2001601_02.jpg", "2001601_03.jpg", "2001601_04.jpg", "2001601_05.jpg", "2001601_06.jpg", "2001601_07.jpg", "2001601_08.jpg", "2001601_09.jpg", "2001601_10.jpg", "2001601_11.jpg", "2001601_12.jpg", "2001601_13.jpg", "2001601_14.jpg", "2001601_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 \"house with house rider on a grass ground\",\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: 1\n factual consistency: 1\n" } ]

[ "3002434_00.jpg", "3002434_01.jpg", "3002434_02.jpg", "3002434_03.jpg", "3002434_04.jpg", "3002434_05.jpg", "3002434_06.jpg", "3002434_07.jpg", "3002434_08.jpg", "3002434_09.jpg", "3002434_10.jpg", "3002434_11.jpg", "3002434_12.jpg", "3002434_13.jpg", "3002434_14.jpg", "3002434_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 photorealistic cinematic scene shot by Steven Spielberg of the inside of a cave tunnel with a 2feet deep 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> " }, { "from": "gpt", "value": "visual quality: 3\n temporal consistency: 3\n dynamic degree: 2\n text-to-video alignment: 3\n factual consistency: 3\n" } ]

[ "2001303_00.jpg", "2001303_01.jpg", "2001303_02.jpg", "2001303_03.jpg", "2001303_04.jpg", "2001303_05.jpg", "2001303_06.jpg", "2001303_07.jpg", "2001303_08.jpg", "2001303_09.jpg", "2001303_10.jpg", "2001303_11.jpg", "2001303_12.jpg", "2001303_13.jpg", "2001303_14.jpg", "2001303_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 learned elderly black woman from walking in london beckoning the viewer\",\nall the frames of video are as follows:\n\n<image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> " }, { "from": "gpt", "value": "visual quality: 3\n temporal consistency: 3\n dynamic degree: 3\n text-to-video alignment: 3\n factual consistency: 3\n" } ]

[ "7000161_00.jpg", "7000161_01.jpg", "7000161_02.jpg", "7000161_03.jpg", "7000161_04.jpg", "7000161_05.jpg", "7000161_06.jpg", "7000161_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 crowd of thousands of happy danes with golden crowns, golden crowns falling from the sky, confetti and balloon\",\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: 2\n factual consistency: 1\n" } ]

[ "6001204_00.jpg", "6001204_01.jpg", "6001204_02.jpg", "6001204_03.jpg", "6001204_04.jpg", "6001204_05.jpg", "6001204_06.jpg", "6001204_07.jpg", "6001204_08.jpg", "6001204_09.jpg", "6001204_10.jpg", "6001204_11.jpg", "6001204_12.jpg", "6001204_13.jpg", "6001204_14.jpg", "6001204_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 \"Describe the bond and friendship between a cat and its owner. 4k, photorealistic, realism, stunning\",\nall the frames of video are as follows:\n\n<image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> " }, { "from": "gpt", "value": "visual quality: 3\n temporal consistency: 2\n dynamic degree: 3\n text-to-video alignment: 3\n factual consistency: 1\n" } ]

[ "2001019_00.jpg", "2001019_01.jpg", "2001019_02.jpg", "2001019_03.jpg", "2001019_04.jpg", "2001019_05.jpg", "2001019_06.jpg", "2001019_07.jpg", "2001019_08.jpg", "2001019_09.jpg", "2001019_10.jpg", "2001019_11.jpg", "2001019_12.jpg", "2001019_13.jpg", "2001019_14.jpg", "2001019_15.jpg" ]

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

[ "1000196_00.jpg", "1000196_01.jpg", "1000196_02.jpg", "1000196_03.jpg", "1000196_04.jpg", "1000196_05.jpg", "1000196_06.jpg", "1000196_07.jpg", "1000196_08.jpg", "1000196_09.jpg", "1000196_10.jpg", "1000196_11.jpg", "1000196_12.jpg", "1000196_13.jpg", "1000196_14.jpg", "1000196_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 standing in meadows , ghibli style , wind is blowing , grass moving , beautiful scene\",\nall the frames of video are as follows:\n\n<image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> " }, { "from": "gpt", "value": "visual quality: 1\n temporal consistency: 2\n dynamic degree: 3\n text-to-video alignment: 2\n factual consistency: 1\n" } ]

[ "0000024_00.jpg", "0000024_01.jpg", "0000024_02.jpg", "0000024_03.jpg", "0000024_04.jpg", "0000024_05.jpg", "0000024_06.jpg", "0000024_07.jpg", "0000024_08.jpg", "0000024_09.jpg", "0000024_10.jpg", "0000024_11.jpg", "0000024_12.jpg", "0000024_13.jpg", "0000024_14.jpg", "0000024_15.jpg", "0000024_16.jpg", "0000024_17.jpg", "0000024_18.jpg", "0000024_19.jpg", "0000024_20.jpg", "0000024_21.jpg", "0000024_22.jpg", "0000024_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 close up of a hand playing a C Major chord on a syntheizer\",\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: 3\n factual consistency: 2\n" } ]

[ "2001039_00.jpg", "2001039_01.jpg", "2001039_02.jpg", "2001039_03.jpg", "2001039_04.jpg", "2001039_05.jpg", "2001039_06.jpg", "2001039_07.jpg", "2001039_08.jpg", "2001039_09.jpg", "2001039_10.jpg", "2001039_11.jpg", "2001039_12.jpg", "2001039_13.jpg", "2001039_14.jpg", "2001039_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 cinematic scene in space of a large spacecraft 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> " }, { "from": "gpt", "value": "visual quality: 3\n temporal consistency: 3\n dynamic degree: 3\n text-to-video alignment: 3\n factual consistency: 1\n" } ]

[ "6001489_00.jpg", "6001489_01.jpg", "6001489_02.jpg", "6001489_03.jpg", "6001489_04.jpg", "6001489_05.jpg", "6001489_06.jpg", "6001489_07.jpg", "6001489_08.jpg", "6001489_09.jpg", "6001489_10.jpg", "6001489_11.jpg", "6001489_12.jpg", "6001489_13.jpg", "6001489_14.jpg", "6001489_15.jpg" ]

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

[ "7001745_00.jpg", "7001745_01.jpg", "7001745_02.jpg", "7001745_03.jpg", "7001745_04.jpg", "7001745_05.jpg", "7001745_06.jpg", "7001745_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 Germany, Wes Anderson style,heavily raining, thunderstorm, well lit, timelapse, hyper realistic, high octane render, high texture, extreme zoom in\",\nall the frames of video are as follows:\n\n<image> <image> <image> <image> <image> <image> <image> <image> " }, { "from": "gpt", "value": "visual quality: 3\n temporal consistency: 3\n dynamic degree: 3\n text-to-video alignment: 2\n factual consistency: 3\n" } ]

[ "2000801_00.jpg", "2000801_01.jpg", "2000801_02.jpg", "2000801_03.jpg", "2000801_04.jpg", "2000801_05.jpg", "2000801_06.jpg", "2000801_07.jpg", "2000801_08.jpg", "2000801_09.jpg", "2000801_10.jpg", "2000801_11.jpg", "2000801_12.jpg", "2000801_13.jpg", "2000801_14.jpg", "2000801_15.jpg" ]

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

[ "2002751_00.jpg", "2002751_01.jpg", "2002751_02.jpg", "2002751_03.jpg", "2002751_04.jpg", "2002751_05.jpg", "2002751_06.jpg", "2002751_07.jpg", "2002751_08.jpg", "2002751_09.jpg", "2002751_10.jpg", "2002751_11.jpg", "2002751_12.jpg", "2002751_13.jpg", "2002751_14.jpg", "2002751_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 \"Extremist groups on both sides sought to undermine any progress toward peace, perpetuating a cycle of bloodshed that stained the land and consumed countless lives.\",\nall the frames of video are as follows:\n\n<image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> " }, { "from": "gpt", "value": "visual quality: 2\n temporal consistency: 3\n dynamic degree: 3\n text-to-video alignment: 1\n factual consistency: 3\n" } ]

[ "5000505_00.jpg", "5000505_01.jpg", "5000505_02.jpg", "5000505_03.jpg", "5000505_04.jpg", "5000505_05.jpg", "5000505_06.jpg", "5000505_07.jpg", "5000505_08.jpg", "5000505_09.jpg", "5000505_10.jpg", "5000505_11.jpg", "5000505_12.jpg", "5000505_13.jpg", "5000505_14.jpg", "5000505_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 \"On the billboard on the right side of the highway guardrail, on the foggy highway, it looks over the car window that says slow motion on the screen, running car pov\",\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: 2\n factual consistency: 3\n" } ]

[ "5000916_00.jpg", "5000916_01.jpg", "5000916_02.jpg", "5000916_03.jpg", "5000916_04.jpg", "5000916_05.jpg", "5000916_06.jpg", "5000916_07.jpg", "5000916_08.jpg", "5000916_09.jpg", "5000916_10.jpg", "5000916_11.jpg", "5000916_12.jpg", "5000916_13.jpg", "5000916_14.jpg", "5000916_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 \"Winter wonderland with cabin in background and Christmas lights on telephone towers\",\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: 4\n temporal consistency: 4\n dynamic degree: 4\n text-to-video alignment: 4\n factual consistency: 4\n" } ]

[ "1000571_00.jpg", "1000571_01.jpg", "1000571_02.jpg", "1000571_03.jpg", "1000571_04.jpg", "1000571_05.jpg", "1000571_06.jpg", "1000571_07.jpg", "1000571_08.jpg", "1000571_09.jpg", "1000571_10.jpg", "1000571_11.jpg", "1000571_12.jpg", "1000571_13.jpg", "1000571_14.jpg", "1000571_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 \"Drone footage of war torn dubai\",\nall the frames of video are as follows:\n\n<image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> " }, { "from": "gpt", "value": "visual quality: 1\n temporal consistency: 1\n dynamic degree: 3\n text-to-video alignment: 2\n factual consistency: 1\n" } ]

[ "4001024_00.jpg", "4001024_01.jpg", "4001024_02.jpg", "4001024_03.jpg", "4001024_04.jpg", "4001024_05.jpg", "4001024_06.jpg", "4001024_07.jpg", "4001024_08.jpg", "4001024_09.jpg", "4001024_10.jpg", "4001024_11.jpg", "4001024_12.jpg", "4001024_13.jpg", "4001024_14.jpg", "4001024_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 image of war. Fumes. Weaponry.\",\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: 1\n factual consistency: 3\n" } ]

[ "2000270_00.jpg", "2000270_01.jpg", "2000270_02.jpg", "2000270_03.jpg", "2000270_04.jpg", "2000270_05.jpg", "2000270_06.jpg", "2000270_07.jpg", "2000270_08.jpg", "2000270_09.jpg", "2000270_10.jpg", "2000270_11.jpg", "2000270_12.jpg", "2000270_13.jpg", "2000270_14.jpg", "2000270_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 hamburger with melt cheese inside. The angle of view has to be close to appreciating the meat with 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: 3\n temporal consistency: 3\n dynamic degree: 3\n text-to-video alignment: 3\n factual consistency: 3\n" } ]

[ "6001044_00.jpg", "6001044_01.jpg", "6001044_02.jpg", "6001044_03.jpg", "6001044_04.jpg", "6001044_05.jpg", "6001044_06.jpg", "6001044_07.jpg", "6001044_08.jpg", "6001044_09.jpg", "6001044_10.jpg", "6001044_11.jpg", "6001044_12.jpg", "6001044_13.jpg", "6001044_14.jpg", "6001044_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 walking in an autumn park, Fallen maple leaves are dancing in autumn.\",\nall the frames of video are as follows:\n\n<image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> " }, { "from": "gpt", "value": "visual quality: 1\n temporal consistency: 2\n dynamic degree: 3\n text-to-video alignment: 3\n factual consistency: 1\n" } ]

[ "7001048_00.jpg", "7001048_01.jpg", "7001048_02.jpg", "7001048_03.jpg", "7001048_04.jpg", "7001048_05.jpg", "7001048_06.jpg", "7001048_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 \"many pschedelic snakes in kaleidoscopic scene 1960s trippy 4k ar 16:9\",\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: 1\n factual consistency: 1\n" } ]

[ "0000590_00.jpg", "0000590_01.jpg", "0000590_02.jpg", "0000590_03.jpg", "0000590_04.jpg", "0000590_05.jpg", "0000590_06.jpg", "0000590_07.jpg", "0000590_08.jpg", "0000590_09.jpg", "0000590_10.jpg", "0000590_11.jpg", "0000590_12.jpg", "0000590_13.jpg", "0000590_14.jpg", "0000590_15.jpg", "0000590_16.jpg", "0000590_17.jpg", "0000590_18.jpg", "0000590_19.jpg", "0000590_20.jpg", "0000590_21.jpg", "0000590_22.jpg", "0000590_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 baby is crawling towards me.\",\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: 3\n factual consistency: 1\n" } ]

[ "2002955_00.jpg", "2002955_01.jpg", "2002955_02.jpg", "2002955_03.jpg", "2002955_04.jpg", "2002955_05.jpg", "2002955_06.jpg", "2002955_07.jpg", "2002955_08.jpg", "2002955_09.jpg", "2002955_10.jpg", "2002955_11.jpg", "2002955_12.jpg", "2002955_13.jpg", "2002955_14.jpg", "2002955_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 \"pink skinned male faerie doing magic, camera zoom out, transformation\",\nall the frames of video are as follows:\n\n<image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> " }, { "from": "gpt", "value": "visual quality: 3\n temporal consistency: 3\n dynamic degree: 3\n text-to-video alignment: 3\n factual consistency: 2\n" } ]

[ "2002562_00.jpg", "2002562_01.jpg", "2002562_02.jpg", "2002562_03.jpg", "2002562_04.jpg", "2002562_05.jpg", "2002562_06.jpg", "2002562_07.jpg", "2002562_08.jpg", "2002562_09.jpg", "2002562_10.jpg", "2002562_11.jpg", "2002562_12.jpg", "2002562_13.jpg", "2002562_14.jpg", "2002562_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 street fading in and out slowly and couple watching it from terrace\",\nall the frames of video are as follows:\n\n<image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> " }, { "from": "gpt", "value": "visual quality: 3\n temporal consistency: 3\n dynamic degree: 2\n text-to-video alignment: 3\n factual consistency: 3\n" } ]

[ "5000695_00.jpg", "5000695_01.jpg", "5000695_02.jpg", "5000695_03.jpg", "5000695_04.jpg", "5000695_05.jpg", "5000695_06.jpg", "5000695_07.jpg", "5000695_08.jpg", "5000695_09.jpg", "5000695_10.jpg", "5000695_11.jpg", "5000695_12.jpg", "5000695_13.jpg", "5000695_14.jpg", "5000695_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 \"so I walked on down past the burger crown\",\nall the frames of video are as follows:\n\n<image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> " }, { "from": "gpt", "value": "visual quality: 3\n temporal consistency: 3\n dynamic degree: 2\n text-to-video alignment: 3\n factual consistency: 3\n" } ]

[ "1002173_00.jpg", "1002173_01.jpg", "1002173_02.jpg", "1002173_03.jpg", "1002173_04.jpg", "1002173_05.jpg", "1002173_06.jpg", "1002173_07.jpg", "1002173_08.jpg", "1002173_09.jpg", "1002173_10.jpg", "1002173_11.jpg", "1002173_12.jpg", "1002173_13.jpg", "1002173_14.jpg", "1002173_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 \"Meeting between the journalist and a mysterious figure, their faces partially obscured by shadows\",\nall the frames of video are as follows:\n\n<image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> " }, { "from": "gpt", "value": "visual quality: 1\n temporal consistency: 2\n dynamic degree: 2\n text-to-video alignment: 2\n factual consistency: 1\n" } ]

[ "2002515_00.jpg", "2002515_01.jpg", "2002515_02.jpg", "2002515_03.jpg", "2002515_04.jpg", "2002515_05.jpg", "2002515_06.jpg", "2002515_07.jpg", "2002515_08.jpg", "2002515_09.jpg", "2002515_10.jpg", "2002515_11.jpg", "2002515_12.jpg", "2002515_13.jpg", "2002515_14.jpg", "2002515_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 king of darkness sitting on his throne\",\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: 3\n factual consistency: 3\n" } ]

[ "8000218_00.jpg", "8000218_01.jpg", "8000218_02.jpg", "8000218_03.jpg", "8000218_04.jpg", "8000218_05.jpg", "8000218_06.jpg", "8000218_07.jpg", "8000218_08.jpg", "8000218_09.jpg", "8000218_10.jpg", "8000218_11.jpg", "8000218_12.jpg", "8000218_13.jpg", "8000218_14.jpg", "8000218_15.jpg", "8000218_16.jpg", "8000218_17.jpg", "8000218_18.jpg", "8000218_19.jpg", "8000218_20.jpg", "8000218_21.jpg", "8000218_22.jpg", "8000218_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 lady walking in a busy street and peoples are enjoying in the street light are cloudy white balance is 3400 shadow little more pink and tint and camera move 360 degree and moving towards to the subject\",\nall the frames of video are as follows:\n\n<image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> " }, { "from": "gpt", "value": "visual quality: 1\n temporal consistency: 1\n dynamic degree: 3\n text-to-video alignment: 2\n factual consistency: 1\n" } ]

[ "4001838_00.jpg", "4001838_01.jpg", "4001838_02.jpg", "4001838_03.jpg", "4001838_04.jpg", "4001838_05.jpg", "4001838_06.jpg", "4001838_07.jpg", "4001838_08.jpg", "4001838_09.jpg", "4001838_10.jpg", "4001838_11.jpg", "4001838_12.jpg", "4001838_13.jpg", "4001838_14.jpg", "4001838_15.jpg" ]

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

[ "4001223_00.jpg", "4001223_01.jpg", "4001223_02.jpg", "4001223_03.jpg", "4001223_04.jpg", "4001223_05.jpg", "4001223_06.jpg", "4001223_07.jpg", "4001223_08.jpg", "4001223_09.jpg", "4001223_10.jpg", "4001223_11.jpg", "4001223_12.jpg", "4001223_13.jpg", "4001223_14.jpg", "4001223_15.jpg" ]

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

[ "6001558_00.jpg", "6001558_01.jpg", "6001558_02.jpg", "6001558_03.jpg", "6001558_04.jpg", "6001558_05.jpg", "6001558_06.jpg", "6001558_07.jpg", "6001558_08.jpg", "6001558_09.jpg", "6001558_10.jpg", "6001558_11.jpg", "6001558_12.jpg", "6001558_13.jpg", "6001558_14.jpg", "6001558_15.jpg" ]

[ { "from": "human", "value": "\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos,\nplease watch the following frames of a given video and see the text prompt for generating the video,\nthen give scores from 5 different dimensions:\n(1) visual quality: the quality of the video in terms of clearness, resolution, brightness, and color\n(2) temporal consistency, both the consistency of objects or humans and the smoothness of motion or movements\n(3) dynamic degree, the degree of dynamic changes\n(4) text-to-video alignment, the alignment between the text prompt and the video content\n(5) factual consistency, the consistency of the video content with the common-sense and factual knowledge\n\nfor each dimension, output a float number from 1.0 to 4.0,\nthe higher the number is, the better the video performs in that sub-score, \nthe lowest 1.0 means Bad, the highest 4.0 means Perfect/Real (the video is like a real video)\nHere is an output example:\nvisual quality: 3.2\ntemporal consistency: 2.7\ndynamic degree: 4.0\ntext-to-video alignment: 2.3\nfactual consistency: 1.8\n\nFor this video, the text prompt is \"two people having a chat over coffee motion:0\",\nall the frames of video are as follows:\n\n<image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> " }, { "from": "gpt", "value": "visual quality: 1\n temporal consistency: 2\n dynamic degree: 3\n text-to-video alignment: 3\n factual consistency: 1\n" } ]

[ "7000411_00.jpg", "7000411_01.jpg", "7000411_02.jpg", "7000411_03.jpg", "7000411_04.jpg", "7000411_05.jpg", "7000411_06.jpg", "7000411_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 \"what a bottle of alcohol will look like in the future\",\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: 3\n factual consistency: 3\n" } ]

[ "0000283_00.jpg", "0000283_01.jpg", "0000283_02.jpg", "0000283_03.jpg", "0000283_04.jpg", "0000283_05.jpg", "0000283_06.jpg", "0000283_07.jpg", "0000283_08.jpg", "0000283_09.jpg", "0000283_10.jpg", "0000283_11.jpg", "0000283_12.jpg", "0000283_13.jpg", "0000283_14.jpg", "0000283_15.jpg", "0000283_16.jpg", "0000283_17.jpg", "0000283_18.jpg", "0000283_19.jpg", "0000283_20.jpg", "0000283_21.jpg", "0000283_22.jpg", "0000283_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 \"sun exploding in the sky, dave gibbons 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: 3\n temporal consistency: 1\n dynamic degree: 1\n text-to-video alignment: 3\n factual consistency: 3\n" } ]

[ "1001096_00.jpg", "1001096_01.jpg", "1001096_02.jpg", "1001096_03.jpg", "1001096_04.jpg", "1001096_05.jpg", "1001096_06.jpg", "1001096_07.jpg", "1001096_08.jpg", "1001096_09.jpg", "1001096_10.jpg", "1001096_11.jpg", "1001096_12.jpg", "1001096_13.jpg", "1001096_14.jpg", "1001096_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 \"james hetfield from metallica, longh hair, sitting alone playing guitar, christmas tree, black and white 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> " }, { "from": "gpt", "value": "visual quality: 1\n temporal consistency: 1\n dynamic degree: 3\n text-to-video alignment: 2\n factual consistency: 1\n" } ]

[ "0001886_00.jpg", "0001886_01.jpg", "0001886_02.jpg", "0001886_03.jpg", "0001886_04.jpg", "0001886_05.jpg", "0001886_06.jpg", "0001886_07.jpg", "0001886_08.jpg", "0001886_09.jpg", "0001886_10.jpg", "0001886_11.jpg", "0001886_12.jpg", "0001886_13.jpg", "0001886_14.jpg", "0001886_15.jpg", "0001886_16.jpg", "0001886_17.jpg", "0001886_18.jpg", "0001886_19.jpg", "0001886_20.jpg", "0001886_21.jpg", "0001886_22.jpg", "0001886_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 \"October is world mental health awareness month\",\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: 1\n text-to-video alignment: 1\n factual consistency: 3\n" } ]

[ "7001195_00.jpg", "7001195_01.jpg", "7001195_02.jpg", "7001195_03.jpg", "7001195_04.jpg", "7001195_05.jpg", "7001195_06.jpg", "7001195_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 \"an enormous white sphere , realistic , monochrome, cgi, dramatic, negative space\",\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: 2\n text-to-video alignment: 2\n factual consistency: 3\n" } ]

[ "0001834_00.jpg", "0001834_01.jpg", "0001834_02.jpg", "0001834_03.jpg", "0001834_04.jpg", "0001834_05.jpg", "0001834_06.jpg", "0001834_07.jpg", "0001834_08.jpg", "0001834_09.jpg", "0001834_10.jpg", "0001834_11.jpg", "0001834_12.jpg", "0001834_13.jpg", "0001834_14.jpg", "0001834_15.jpg", "0001834_16.jpg", "0001834_17.jpg", "0001834_18.jpg", "0001834_19.jpg", "0001834_20.jpg", "0001834_21.jpg", "0001834_22.jpg", "0001834_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 \"Neon minimalism in style of techno futuristic, luminous 3D objects, Carnelian, man studio photography, Pop up editorial photography,\",\nall the frames of video are as follows:\n\n<image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> " }, { "from": "gpt", "value": "visual quality: 3\n temporal consistency: 3\n dynamic degree: 1\n text-to-video alignment: 3\n factual consistency: 3\n" } ]

[ "7001031_00.jpg", "7001031_01.jpg", "7001031_02.jpg", "7001031_03.jpg", "7001031_04.jpg", "7001031_05.jpg", "7001031_06.jpg", "7001031_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 big wooden clock in an abandoned castle where the camera comes from the door of the room\",\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: 3\n factual consistency: 3\n" } ]

[ "2001653_00.jpg", "2001653_01.jpg", "2001653_02.jpg", "2001653_03.jpg", "2001653_04.jpg", "2001653_05.jpg", "2001653_06.jpg", "2001653_07.jpg", "2001653_08.jpg", "2001653_09.jpg", "2001653_10.jpg", "2001653_11.jpg", "2001653_12.jpg", "2001653_13.jpg", "2001653_14.jpg", "2001653_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 \"man going to travel by plane\",\nall the frames of video are as follows:\n\n<image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> " }, { "from": "gpt", "value": "visual quality: 3\n temporal consistency: 2\n dynamic degree: 3\n text-to-video alignment: 3\n factual consistency: 1\n" } ]

[ "8000161_00.jpg", "8000161_01.jpg", "8000161_02.jpg", "8000161_03.jpg", "8000161_04.jpg", "8000161_05.jpg", "8000161_06.jpg", "8000161_07.jpg", "8000161_08.jpg", "8000161_09.jpg", "8000161_10.jpg", "8000161_11.jpg", "8000161_12.jpg", "8000161_13.jpg", "8000161_14.jpg", "8000161_15.jpg", "8000161_16.jpg", "8000161_17.jpg", "8000161_18.jpg", "8000161_19.jpg", "8000161_20.jpg", "8000161_21.jpg", "8000161_22.jpg", "8000161_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 \"Lifestyle photography photo of a young pretty Japanese woman, with long, black, straight hair, pouring tea in a traditional tea ceremony, looking at viewer, upper body framing, in a Japanese tea room, symmetrical background, under (natural lighting:1.3), shot at eye level, on a Sony A7111, with a (soft focus filter:1.3),(in the style of Wes Anderson:1.3)\",\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: 2\n factual consistency: 1\n" } ]

[ "5000573_00.jpg", "5000573_01.jpg", "5000573_02.jpg", "5000573_03.jpg", "5000573_04.jpg", "5000573_05.jpg", "5000573_06.jpg", "5000573_07.jpg", "5000573_08.jpg", "5000573_09.jpg", "5000573_10.jpg", "5000573_11.jpg", "5000573_12.jpg", "5000573_13.jpg", "5000573_14.jpg", "5000573_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 goat flies over the fire in the haunted 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: 2\n temporal consistency: 2\n dynamic degree: 2\n text-to-video alignment: 1\n factual consistency: 2\n" } ]

[ "2001530_00.jpg", "2001530_01.jpg", "2001530_02.jpg", "2001530_03.jpg", "2001530_04.jpg", "2001530_05.jpg", "2001530_06.jpg", "2001530_07.jpg", "2001530_08.jpg", "2001530_09.jpg", "2001530_10.jpg", "2001530_11.jpg", "2001530_12.jpg", "2001530_13.jpg", "2001530_14.jpg", "2001530_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 \"It’s even been proposed that it could be a relic of a forgotten culture that existed when sea levels were significantly lower than today.\",\nall the frames of video are as follows:\n\n<image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> " }, { "from": "gpt", "value": "visual quality: 3\n temporal consistency: 3\n dynamic degree: 3\n text-to-video alignment: 3\n factual consistency: 3\n" } ]

[ "1002703_00.jpg", "1002703_01.jpg", "1002703_02.jpg", "1002703_03.jpg", "1002703_04.jpg", "1002703_05.jpg", "1002703_06.jpg", "1002703_07.jpg", "1002703_08.jpg", "1002703_09.jpg", "1002703_10.jpg", "1002703_11.jpg", "1002703_12.jpg", "1002703_13.jpg", "1002703_14.jpg", "1002703_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 \"With a determined heart, lion set off on his journey. He padded through the tall grass, his tiny paws barely making a sound. The world beyond the meadow was vast and filled with wonders he had never seen before\",\nall the frames of video are as follows:\n\n<image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> " }, { "from": "gpt", "value": "visual quality: 1\n temporal consistency: 2\n dynamic degree: 3\n text-to-video alignment: 2\n factual consistency: 1\n" } ]

[ "3000981_00.jpg", "3000981_01.jpg", "3000981_02.jpg", "3000981_03.jpg", "3000981_04.jpg", "3000981_05.jpg", "3000981_06.jpg", "3000981_07.jpg", "3000981_08.jpg", "3000981_09.jpg", "3000981_10.jpg", "3000981_11.jpg", "3000981_12.jpg", "3000981_13.jpg", "3000981_14.jpg", "3000981_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 men climb one montain with green hat\",\nall the frames of video are as follows:\n\n<image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> " }, { "from": "gpt", "value": "visual quality: 2\n temporal consistency: 1\n dynamic degree: 3\n text-to-video alignment: 1\n factual consistency: 1\n" } ]

[ "6001430_00.jpg", "6001430_01.jpg", "6001430_02.jpg", "6001430_03.jpg", "6001430_04.jpg", "6001430_05.jpg", "6001430_06.jpg", "6001430_07.jpg", "6001430_08.jpg", "6001430_09.jpg", "6001430_10.jpg", "6001430_11.jpg", "6001430_12.jpg", "6001430_13.jpg", "6001430_14.jpg", "6001430_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 super hero persian carpet design walking in the 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: 2\n dynamic degree: 3\n text-to-video alignment: 2\n factual consistency: 1\n" } ]

[ "1002776_00.jpg", "1002776_01.jpg", "1002776_02.jpg", "1002776_03.jpg", "1002776_04.jpg", "1002776_05.jpg", "1002776_06.jpg", "1002776_07.jpg", "1002776_08.jpg", "1002776_09.jpg", "1002776_10.jpg", "1002776_11.jpg", "1002776_12.jpg", "1002776_13.jpg", "1002776_14.jpg", "1002776_15.jpg" ]

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

[ "1001920_00.jpg", "1001920_01.jpg", "1001920_02.jpg", "1001920_03.jpg", "1001920_04.jpg", "1001920_05.jpg", "1001920_06.jpg", "1001920_07.jpg", "1001920_08.jpg", "1001920_09.jpg", "1001920_10.jpg", "1001920_11.jpg", "1001920_12.jpg", "1001920_13.jpg", "1001920_14.jpg", "1001920_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 \"producing fertilisers with your clothes and gallons written on them InnovaLogicLAB\",\nall the frames of video are as follows:\n\n<image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> " }, { "from": "gpt", "value": "visual quality: 1\n temporal consistency: 1\n dynamic degree: 3\n text-to-video alignment: 2\n factual consistency: 1\n" } ]

[ "1001734_00.jpg", "1001734_01.jpg", "1001734_02.jpg", "1001734_03.jpg", "1001734_04.jpg", "1001734_05.jpg", "1001734_06.jpg", "1001734_07.jpg", "1001734_08.jpg", "1001734_09.jpg", "1001734_10.jpg", "1001734_11.jpg", "1001734_12.jpg", "1001734_13.jpg", "1001734_14.jpg", "1001734_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 statue in an empty field that begins to crumble\",\nall the frames of video are as follows:\n\n<image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> " }, { "from": "gpt", "value": "visual quality: 1\n temporal consistency: 1\n dynamic degree: 3\n text-to-video alignment: 2\n factual consistency: 1\n" } ]

[ "1000577_00.jpg", "1000577_01.jpg", "1000577_02.jpg", "1000577_03.jpg", "1000577_04.jpg", "1000577_05.jpg", "1000577_06.jpg", "1000577_07.jpg", "1000577_08.jpg", "1000577_09.jpg", "1000577_10.jpg", "1000577_11.jpg", "1000577_12.jpg", "1000577_13.jpg", "1000577_14.jpg", "1000577_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 \"white muscular man showing of his muscle like escanor of seven deadly sins anime in an Winter storm in mount everest, 4k , cinematic lighting , digital art , photorealistic\",\nall the frames of video are as follows:\n\n<image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> " }, { "from": "gpt", "value": "visual quality: 1\n temporal consistency: 1\n dynamic degree: 3\n text-to-video alignment: 2\n factual consistency: 1\n" } ]

[ "8000155_00.jpg", "8000155_01.jpg", "8000155_02.jpg", "8000155_03.jpg", "8000155_04.jpg", "8000155_05.jpg", "8000155_06.jpg", "8000155_07.jpg", "8000155_08.jpg", "8000155_09.jpg", "8000155_10.jpg", "8000155_11.jpg", "8000155_12.jpg", "8000155_13.jpg", "8000155_14.jpg", "8000155_15.jpg", "8000155_16.jpg", "8000155_17.jpg", "8000155_18.jpg", "8000155_19.jpg", "8000155_20.jpg", "8000155_21.jpg", "8000155_22.jpg", "8000155_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 \"nighttime, Jimi Hendrix appears as an aperition and sheds psychedelic fairydust from his guitar as floats above us .\",\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: 2\n factual consistency: 1\n" } ]

[ "2002449_00.jpg", "2002449_01.jpg", "2002449_02.jpg", "2002449_03.jpg", "2002449_04.jpg", "2002449_05.jpg", "2002449_06.jpg", "2002449_07.jpg", "2002449_08.jpg", "2002449_09.jpg", "2002449_10.jpg", "2002449_11.jpg", "2002449_12.jpg", "2002449_13.jpg", "2002449_14.jpg", "2002449_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 \"cat turning the steering wheel in a sports car going fast\",\nall the frames of video are as follows:\n\n<image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> " }, { "from": "gpt", "value": "visual quality: 3\n temporal consistency: 2\n dynamic degree: 3\n text-to-video alignment: 3\n factual consistency: 1\n" } ]

[ "0002146_00.jpg", "0002146_01.jpg", "0002146_02.jpg", "0002146_03.jpg", "0002146_04.jpg", "0002146_05.jpg", "0002146_06.jpg", "0002146_07.jpg", "0002146_08.jpg", "0002146_09.jpg", "0002146_10.jpg", "0002146_11.jpg", "0002146_12.jpg", "0002146_13.jpg", "0002146_14.jpg", "0002146_15.jpg", "0002146_16.jpg", "0002146_17.jpg", "0002146_18.jpg", "0002146_19.jpg", "0002146_20.jpg", "0002146_21.jpg", "0002146_22.jpg", "0002146_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 \"Illustrate the Snow Queen hosting a grand winter ball in her ice palace.\",\nall the frames of video are as follows:\n\n<image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> " }, { "from": "gpt", "value": "visual quality: 3\n temporal consistency: 3\n dynamic degree: 1\n text-to-video alignment: 3\n factual consistency: 3\n" } ]

[ "3000683_00.jpg", "3000683_01.jpg", "3000683_02.jpg", "3000683_03.jpg", "3000683_04.jpg", "3000683_05.jpg", "3000683_06.jpg", "3000683_07.jpg", "3000683_08.jpg", "3000683_09.jpg", "3000683_10.jpg", "3000683_11.jpg", "3000683_12.jpg", "3000683_13.jpg", "3000683_14.jpg", "3000683_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 old wizard in a blue cloak fights with other wizards in the forest in winter\",\nall the frames of video are as follows:\n\n<image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> " }, { "from": "gpt", "value": "visual quality: 2\n temporal consistency: 1\n dynamic degree: 3\n text-to-video alignment: 1\n factual consistency: 1\n" } ]

[ "7000506_00.jpg", "7000506_01.jpg", "7000506_02.jpg", "7000506_03.jpg", "7000506_04.jpg", "7000506_05.jpg", "7000506_06.jpg", "7000506_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 \"Native americans tribe in an island around a big fire, 70s dark fantasy style, live action, realistic, cinematic, vintage, HD realistic, ULTRA HD, 70s movie style\",\nall the frames of video are as follows:\n\n<image> <image> <image> <image> <image> <image> <image> <image> " }, { "from": "gpt", "value": "visual quality: 2\n temporal consistency: 3\n dynamic degree: 3\n text-to-video alignment: 3\n factual consistency: 3\n" } ]

[ "8000643_00.jpg", "8000643_01.jpg", "8000643_02.jpg", "8000643_03.jpg", "8000643_04.jpg", "8000643_05.jpg", "8000643_06.jpg", "8000643_07.jpg", "8000643_08.jpg", "8000643_09.jpg", "8000643_10.jpg", "8000643_11.jpg", "8000643_12.jpg", "8000643_13.jpg", "8000643_14.jpg", "8000643_15.jpg", "8000643_16.jpg", "8000643_17.jpg", "8000643_18.jpg", "8000643_19.jpg", "8000643_20.jpg", "8000643_21.jpg", "8000643_22.jpg", "8000643_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 \"draw a anime animated chain flying pass camera, close 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> <image> <image> <image> <image> <image> <image> <image> <image> " }, { "from": "gpt", "value": "visual quality: 1\n temporal consistency: 2\n dynamic degree: 3\n text-to-video alignment: 1\n factual consistency: 1\n" } ]

[ "6001140_00.jpg", "6001140_01.jpg", "6001140_02.jpg", "6001140_03.jpg", "6001140_04.jpg", "6001140_05.jpg", "6001140_06.jpg", "6001140_07.jpg", "6001140_08.jpg", "6001140_09.jpg", "6001140_10.jpg", "6001140_11.jpg", "6001140_12.jpg", "6001140_13.jpg", "6001140_14.jpg", "6001140_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 \"firing logo in amazon with the name of akhond\",\nall the frames of video are as follows:\n\n<image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> " }, { "from": "gpt", "value": "visual quality: 1\n temporal consistency: 2\n dynamic degree: 3\n text-to-video alignment: 2\n factual consistency: 1\n" } ]

[ "2000613_00.jpg", "2000613_01.jpg", "2000613_02.jpg", "2000613_03.jpg", "2000613_04.jpg", "2000613_05.jpg", "2000613_06.jpg", "2000613_07.jpg", "2000613_08.jpg", "2000613_09.jpg", "2000613_10.jpg", "2000613_11.jpg", "2000613_12.jpg", "2000613_13.jpg", "2000613_14.jpg", "2000613_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 \"modern comic book illustration style of the outside of the White 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: 1\n text-to-video alignment: 3\n factual consistency: 3\n" } ]

[ "1002008_00.jpg", "1002008_01.jpg", "1002008_02.jpg", "1002008_03.jpg", "1002008_04.jpg", "1002008_05.jpg", "1002008_06.jpg", "1002008_07.jpg", "1002008_08.jpg", "1002008_09.jpg", "1002008_10.jpg", "1002008_11.jpg", "1002008_12.jpg", "1002008_13.jpg", "1002008_14.jpg", "1002008_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 \"Person acting as a frog 30 second HD video\",\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: 1\n factual consistency: 1\n" } ]

[ "7000039_00.jpg", "7000039_01.jpg", "7000039_02.jpg", "7000039_03.jpg", "7000039_04.jpg", "7000039_05.jpg", "7000039_06.jpg", "7000039_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 \"one monge ,in mountain,snow, 4k ,realism\",\nall the frames of video are as follows:\n\n<image> <image> <image> <image> <image> <image> <image> <image> " }, { "from": "gpt", "value": "visual quality: 3\n temporal consistency: 3\n dynamic degree: 3\n text-to-video alignment: 2\n factual consistency: 3\n" } ]

[ "1002242_00.jpg", "1002242_01.jpg", "1002242_02.jpg", "1002242_03.jpg", "1002242_04.jpg", "1002242_05.jpg", "1002242_06.jpg", "1002242_07.jpg", "1002242_08.jpg", "1002242_09.jpg", "1002242_10.jpg", "1002242_11.jpg", "1002242_12.jpg", "1002242_13.jpg", "1002242_14.jpg", "1002242_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 thousand people interacting with each other using futuristic neuralink interfaces, beautiful, inspiring, 4k, magic hour\",\nall the frames of video are as follows:\n\n<image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> " }, { "from": "gpt", "value": "visual quality: 1\n temporal consistency: 1\n dynamic degree: 3\n text-to-video alignment: 2\n factual consistency: 1\n" } ]

[ "2000554_00.jpg", "2000554_01.jpg", "2000554_02.jpg", "2000554_03.jpg", "2000554_04.jpg", "2000554_05.jpg", "2000554_06.jpg", "2000554_07.jpg", "2000554_08.jpg", "2000554_09.jpg", "2000554_10.jpg", "2000554_11.jpg", "2000554_12.jpg", "2000554_13.jpg", "2000554_14.jpg", "2000554_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 \"Narrator: The world watched as Earth made contact with an advanced alien civilization.\",\nall the frames of video are as follows:\n\n<image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> " }, { "from": "gpt", "value": "visual quality: 3\n temporal consistency: 3\n dynamic degree: 3\n text-to-video alignment: 3\n factual consistency: 2\n" } ]

[ "8000792_00.jpg", "8000792_01.jpg", "8000792_02.jpg", "8000792_03.jpg", "8000792_04.jpg", "8000792_05.jpg", "8000792_06.jpg", "8000792_07.jpg", "8000792_08.jpg", "8000792_09.jpg", "8000792_10.jpg", "8000792_11.jpg", "8000792_12.jpg", "8000792_13.jpg", "8000792_14.jpg", "8000792_15.jpg", "8000792_16.jpg", "8000792_17.jpg", "8000792_18.jpg", "8000792_19.jpg", "8000792_20.jpg", "8000792_21.jpg", "8000792_22.jpg", "8000792_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 \"timelapse of a baby peacock hatching from an egg.\",\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: 2\n factual consistency: 1\n" } ]

[ "2000868_00.jpg", "2000868_01.jpg", "2000868_02.jpg", "2000868_03.jpg", "2000868_04.jpg", "2000868_05.jpg", "2000868_06.jpg", "2000868_07.jpg", "2000868_08.jpg", "2000868_09.jpg", "2000868_10.jpg", "2000868_11.jpg", "2000868_12.jpg", "2000868_13.jpg", "2000868_14.jpg", "2000868_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 \"robot lags and hands with body dance\",\nall the frames of video are as follows:\n\n<image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> " }, { "from": "gpt", "value": "visual quality: 3\n temporal consistency: 3\n dynamic degree: 3\n text-to-video alignment: 3\n factual consistency: 3\n" } ]

[ "6001580_00.jpg", "6001580_01.jpg", "6001580_02.jpg", "6001580_03.jpg", "6001580_04.jpg", "6001580_05.jpg", "6001580_06.jpg", "6001580_07.jpg", "6001580_08.jpg", "6001580_09.jpg", "6001580_10.jpg", "6001580_11.jpg", "6001580_12.jpg", "6001580_13.jpg", "6001580_14.jpg", "6001580_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 \"Mountain bike riding in the snow\",\nall the frames of video are as follows:\n\n<image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> " }, { "from": "gpt", "value": "visual quality: 1\n temporal consistency: 2\n dynamic degree: 3\n text-to-video alignment: 3\n factual consistency: 1\n" } ]

[ "0000806_00.jpg", "0000806_01.jpg", "0000806_02.jpg", "0000806_03.jpg", "0000806_04.jpg", "0000806_05.jpg", "0000806_06.jpg", "0000806_07.jpg", "0000806_08.jpg", "0000806_09.jpg", "0000806_10.jpg", "0000806_11.jpg", "0000806_12.jpg", "0000806_13.jpg", "0000806_14.jpg", "0000806_15.jpg", "0000806_16.jpg", "0000806_17.jpg", "0000806_18.jpg", "0000806_19.jpg", "0000806_20.jpg", "0000806_21.jpg", "0000806_22.jpg", "0000806_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 \"Pablo oscobar with Indian old men\",\nall the frames of video are as follows:\n\n<image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> " }, { "from": "gpt", "value": "visual quality: 3\n temporal consistency: 3\n dynamic degree: 2\n text-to-video alignment: 3\n factual consistency: 3\n" } ]

[ "3002380_00.jpg", "3002380_01.jpg", "3002380_02.jpg", "3002380_03.jpg", "3002380_04.jpg", "3002380_05.jpg", "3002380_06.jpg", "3002380_07.jpg", "3002380_08.jpg", "3002380_09.jpg", "3002380_10.jpg", "3002380_11.jpg", "3002380_12.jpg", "3002380_13.jpg", "3002380_14.jpg", "3002380_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 image of a somber Lincoln, sitting alone in a dimly lit room, with a portrait of his late wife, Mary Todd Lincoln, on the wall.\",\nall the frames of video are as follows:\n\n<image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> " }, { "from": "gpt", "value": "visual quality: 2\n temporal consistency: 3\n dynamic degree: 1\n text-to-video alignment: 2\n factual consistency: 2\n" } ]

[ "7001267_00.jpg", "7001267_01.jpg", "7001267_02.jpg", "7001267_03.jpg", "7001267_04.jpg", "7001267_05.jpg", "7001267_06.jpg", "7001267_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 \"realistic fireworks in the background and a lovely couple to show on that\",\nall the frames of video are as follows:\n\n<image> <image> <image> <image> <image> <image> <image> <image> " }, { "from": "gpt", "value": "visual quality: 2\n temporal consistency: 2\n dynamic degree: 3\n text-to-video alignment: 2\n factual consistency: 3\n" } ]

[ "1000376_00.jpg", "1000376_01.jpg", "1000376_02.jpg", "1000376_03.jpg", "1000376_04.jpg", "1000376_05.jpg", "1000376_06.jpg", "1000376_07.jpg", "1000376_08.jpg", "1000376_09.jpg", "1000376_10.jpg", "1000376_11.jpg", "1000376_12.jpg", "1000376_13.jpg", "1000376_14.jpg", "1000376_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 \"video of aliens laughing with humans in Paris\",\nall the frames of video are as follows:\n\n<image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> " }, { "from": "gpt", "value": "visual quality: 1\n temporal consistency: 1\n dynamic degree: 3\n text-to-video alignment: 2\n factual consistency: 1\n" } ]

[ "6000829_00.jpg", "6000829_01.jpg", "6000829_02.jpg", "6000829_03.jpg", "6000829_04.jpg", "6000829_05.jpg", "6000829_06.jpg", "6000829_07.jpg", "6000829_08.jpg", "6000829_09.jpg", "6000829_10.jpg", "6000829_11.jpg", "6000829_12.jpg", "6000829_13.jpg", "6000829_14.jpg", "6000829_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 \"He took out food from the bag and handed the container to the monkey,\",\nall the frames of video are as follows:\n\n<image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> " }, { "from": "gpt", "value": "visual quality: 1\n temporal consistency: 2\n dynamic degree: 3\n text-to-video alignment: 2\n factual consistency: 1\n" } ]

[ "2000951_00.jpg", "2000951_01.jpg", "2000951_02.jpg", "2000951_03.jpg", "2000951_04.jpg", "2000951_05.jpg", "2000951_06.jpg", "2000951_07.jpg", "2000951_08.jpg", "2000951_09.jpg", "2000951_10.jpg", "2000951_11.jpg", "2000951_12.jpg", "2000951_13.jpg", "2000951_14.jpg", "2000951_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 \"rama in ramayan walking towards a dense forest with his brother laxman\",\nall the frames of video are as follows:\n\n<image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> " }, { "from": "gpt", "value": "visual quality: 2\n temporal consistency: 3\n dynamic degree: 3\n text-to-video alignment: 3\n factual consistency: 3\n" } ]

[ "8000640_00.jpg", "8000640_01.jpg", "8000640_02.jpg", "8000640_03.jpg", "8000640_04.jpg", "8000640_05.jpg", "8000640_06.jpg", "8000640_07.jpg", "8000640_08.jpg", "8000640_09.jpg", "8000640_10.jpg", "8000640_11.jpg", "8000640_12.jpg", "8000640_13.jpg", "8000640_14.jpg", "8000640_15.jpg", "8000640_16.jpg", "8000640_17.jpg", "8000640_18.jpg", "8000640_19.jpg", "8000640_20.jpg", "8000640_21.jpg", "8000640_22.jpg", "8000640_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 \"lighting in the middle of an ocean\",\nall the frames of video are as follows:\n\n<image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> " }, { "from": "gpt", "value": "visual quality: 1\n temporal consistency: 2\n dynamic degree: 3\n text-to-video alignment: 2\n factual consistency: 1\n" } ]

[ "8000499_00.jpg", "8000499_01.jpg", "8000499_02.jpg", "8000499_03.jpg", "8000499_04.jpg", "8000499_05.jpg", "8000499_06.jpg", "8000499_07.jpg", "8000499_08.jpg", "8000499_09.jpg", "8000499_10.jpg", "8000499_11.jpg", "8000499_12.jpg", "8000499_13.jpg", "8000499_14.jpg", "8000499_15.jpg", "8000499_16.jpg", "8000499_17.jpg", "8000499_18.jpg", "8000499_19.jpg", "8000499_20.jpg", "8000499_21.jpg", "8000499_22.jpg", "8000499_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 \"digital light show, animate the lights, Burj khalifa 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: 2\n dynamic degree: 3\n text-to-video alignment: 2\n factual consistency: 1\n" } ]

[ "4001950_00.jpg", "4001950_01.jpg", "4001950_02.jpg", "4001950_03.jpg", "4001950_04.jpg", "4001950_05.jpg", "4001950_06.jpg", "4001950_07.jpg", "4001950_08.jpg", "4001950_09.jpg", "4001950_10.jpg", "4001950_11.jpg", "4001950_12.jpg", "4001950_13.jpg", "4001950_14.jpg", "4001950_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 Fast moving Boat with a Man On it in Ocean 16:9 of 60 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: 3\n temporal consistency: 3\n dynamic degree: 3\n text-to-video alignment: 2\n factual consistency: 3\n" } ]