upload subset_20k/test_100 's json, videos and frames

data/subset_20k/test_100/data_insf.json

@@ -0,0 +1,272 @@
+[
+    {
+        "id": "a701934",
+        "images": [
+            "a701934_00.jpg",
+            "a701934_01.jpg",
+            "a701934_02.jpg",
+            "a701934_03.jpg",
+            "a701934_04.jpg",
+            "a701934_05.jpg",
+            "a701934_06.jpg",
+            "a701934_07.jpg"
+        ],
+        "conversations": [
+            {
+                "from": "human",
+                "value": "\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos, \nplease watch the following frames of a given video and see the text prompt for generating the video, \nthen give scores from 7 different dimensions:\n(1) visual quality, \n(2) object consistency,\n(3) dynamic degree,\n(4) motion smoothness,\n(5) text-to-video alignment,\n(6) factual consistency, \n(7) overall score\nfor each dimension, output a number from [1,2,3], in which '1' stands for 'Bad', '2' stands for 'Average', '3' stands for 'Good'.\nHere is an output example: \nvisual quality: 3\nobject consistency: 2 \ndynamic degree: 2\nmotion smoothness: 1\ntext-to-video alignment: 1\nfactual consistency: 2\noverall score: 1\n\nFor this item, the text prompt is chota bheem throws ironman with his barehands  ,\nall the frames of video are as follows: \n\n<image> <image> <image> <image> <image> <image> <image> <image> \n"
+            },
+            {
+                "from": "gpt",
+                "value": "visual quality: 1\n object consistency: 1\n dynamic degree: 3\n motion smoothness: 1\n text-to-video alignment: 1\n factual consistency: 1\n overall score: 1\n"
+            }
+        ]
+    },
+    {
+        "id": "a701052",
+        "images": [
+            "a701052_00.jpg",
+            "a701052_01.jpg",
+            "a701052_02.jpg",
+            "a701052_03.jpg",
+            "a701052_04.jpg",
+            "a701052_05.jpg",
+            "a701052_06.jpg",
+            "a701052_07.jpg"
+        ],
+        "conversations": [
+            {
+                "from": "human",
+                "value": "\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos, \nplease watch the following frames of a given video and see the text prompt for generating the video, \nthen give scores from 7 different dimensions:\n(1) visual quality, \n(2) object consistency,\n(3) dynamic degree,\n(4) motion smoothness,\n(5) text-to-video alignment,\n(6) factual consistency, \n(7) overall score\nfor each dimension, output a number from [1,2,3], in which '1' stands for 'Bad', '2' stands for 'Average', '3' stands for 'Good'.\nHere is an output example: \nvisual quality: 3\nobject consistency: 2 \ndynamic degree: 2\nmotion smoothness: 1\ntext-to-video alignment: 1\nfactual consistency: 2\noverall score: 1\n\nFor this item, the text prompt is Create a captivating animation that begins with a serene landscape of the Snowpeak Mountains. As the wind picks up, simulate the gentle movement of the towering pines, and gradually introduce swirling snowflakes to announce the approaching blizzard. Transition seamlessly to the exterior of a rustic mountain cabin, showcasing its wooden frame and vintage decorations. Use subtle lighting effects to emphasize the coziness inside, with the fire flickering in the stone fireplace. Convey a sense of warmth and anticipation as the characters make their way into the cabin, seeking refuge from the biting wind.  ,\nall the frames of video are as follows: \n\n<image> <image> <image> <image> <image> <image> <image> <image> \n"
+            },
+            {
+                "from": "gpt",
+                "value": "visual quality: 1\n object consistency: 1\n dynamic degree: 3\n motion smoothness: 1\n text-to-video alignment: 1\n factual consistency: 1\n overall score: 1\n"
+            }
+        ]
+    },
+    {
+        "id": "a500395",
+        "images": [
+            "a500395_00.jpg",
+            "a500395_01.jpg",
+            "a500395_02.jpg",
+            "a500395_03.jpg",
+            "a500395_04.jpg",
+            "a500395_05.jpg",
+            "a500395_06.jpg",
+            "a500395_07.jpg",
+            "a500395_08.jpg",
+            "a500395_09.jpg",
+            "a500395_10.jpg",
+            "a500395_11.jpg",
+            "a500395_12.jpg",
+            "a500395_13.jpg",
+            "a500395_14.jpg",
+            "a500395_15.jpg"
+        ],
+        "conversations": [
+            {
+                "from": "human",
+                "value": "\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos, \nplease watch the following frames of a given video and see the text prompt for generating the video, \nthen give scores from 7 different dimensions:\n(1) visual quality, \n(2) object consistency,\n(3) dynamic degree,\n(4) motion smoothness,\n(5) text-to-video alignment,\n(6) factual consistency, \n(7) overall score\nfor each dimension, output a number from [1,2,3], in which '1' stands for 'Bad', '2' stands for 'Average', '3' stands for 'Good'.\nHere is an output example: \nvisual quality: 3\nobject consistency: 2 \ndynamic degree: 2\nmotion smoothness: 1\ntext-to-video alignment: 1\nfactual consistency: 2\noverall score: 1\n\nFor this item, the text prompt is cinematic shot of a drone view of the carribean islands, sun shining  ,\nall the frames of video are as follows: \n\n<image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> \n"
+            },
+            {
+                "from": "gpt",
+                "value": "visual quality: 1\n object consistency: 1\n dynamic degree: 3\n motion smoothness: 1\n text-to-video alignment: 1\n factual consistency: 1\n overall score: 1\n"
+            }
+        ]
+    },
+    {
+        "id": "a701670",
+        "images": [
+            "a701670_00.jpg",
+            "a701670_01.jpg",
+            "a701670_02.jpg",
+            "a701670_03.jpg",
+            "a701670_04.jpg",
+            "a701670_05.jpg",
+            "a701670_06.jpg",
+            "a701670_07.jpg"
+        ],
+        "conversations": [
+            {
+                "from": "human",
+                "value": "\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos, \nplease watch the following frames of a given video and see the text prompt for generating the video, \nthen give scores from 7 different dimensions:\n(1) visual quality, \n(2) object consistency,\n(3) dynamic degree,\n(4) motion smoothness,\n(5) text-to-video alignment,\n(6) factual consistency, \n(7) overall score\nfor each dimension, output a number from [1,2,3], in which '1' stands for 'Bad', '2' stands for 'Average', '3' stands for 'Good'.\nHere is an output example: \nvisual quality: 3\nobject consistency: 2 \ndynamic degree: 2\nmotion smoothness: 1\ntext-to-video alignment: 1\nfactual consistency: 2\noverall score: 1\n\nFor this item, the text prompt is a woman asks her mom to bail her our of prison in torreon, mexico.  ,\nall the frames of video are as follows: \n\n<image> <image> <image> <image> <image> <image> <image> <image> \n"
+            },
+            {
+                "from": "gpt",
+                "value": "visual quality: 1\n object consistency: 1\n dynamic degree: 3\n motion smoothness: 1\n text-to-video alignment: 1\n factual consistency: 1\n overall score: 1\n"
+            }
+        ]
+    },
+    {
+        "id": "a402096",
+        "images": [
+            "a402096_00.jpg",
+            "a402096_01.jpg",
+            "a402096_02.jpg",
+            "a402096_03.jpg",
+            "a402096_04.jpg",
+            "a402096_05.jpg",
+            "a402096_06.jpg",
+            "a402096_07.jpg",
+            "a402096_08.jpg",
+            "a402096_09.jpg",
+            "a402096_10.jpg",
+            "a402096_11.jpg",
+            "a402096_12.jpg",
+            "a402096_13.jpg",
+            "a402096_14.jpg",
+            "a402096_15.jpg"
+        ],
+        "conversations": [
+            {
+                "from": "human",
+                "value": "\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos, \nplease watch the following frames of a given video and see the text prompt for generating the video, \nthen give scores from 7 different dimensions:\n(1) visual quality, \n(2) object consistency,\n(3) dynamic degree,\n(4) motion smoothness,\n(5) text-to-video alignment,\n(6) factual consistency, \n(7) overall score\nfor each dimension, output a number from [1,2,3], in which '1' stands for 'Bad', '2' stands for 'Average', '3' stands for 'Good'.\nHere is an output example: \nvisual quality: 3\nobject consistency: 2 \ndynamic degree: 2\nmotion smoothness: 1\ntext-to-video alignment: 1\nfactual consistency: 2\noverall score: 1\n\nFor this item, the text prompt is In the moonlit darkness of a forgotten town, whispers of a malevolent force echoed through the narrow streets.  ,\nall the frames of video are as follows: \n\n<image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> \n"
+            },
+            {
+                "from": "gpt",
+                "value": "visual quality: 1\n object consistency: 1\n dynamic degree: 3\n motion smoothness: 1\n text-to-video alignment: 1\n factual consistency: 1\n overall score: 1\n"
+            }
+        ]
+    },
+    {
+        "id": "a701180",
+        "images": [
+            "a701180_00.jpg",
+            "a701180_01.jpg",
+            "a701180_02.jpg",
+            "a701180_03.jpg",
+            "a701180_04.jpg",
+            "a701180_05.jpg",
+            "a701180_06.jpg",
+            "a701180_07.jpg"
+        ],
+        "conversations": [
+            {
+                "from": "human",
+                "value": "\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos, \nplease watch the following frames of a given video and see the text prompt for generating the video, \nthen give scores from 7 different dimensions:\n(1) visual quality, \n(2) object consistency,\n(3) dynamic degree,\n(4) motion smoothness,\n(5) text-to-video alignment,\n(6) factual consistency, \n(7) overall score\nfor each dimension, output a number from [1,2,3], in which '1' stands for 'Bad', '2' stands for 'Average', '3' stands for 'Good'.\nHere is an output example: \nvisual quality: 3\nobject consistency: 2 \ndynamic degree: 2\nmotion smoothness: 1\ntext-to-video alignment: 1\nfactual consistency: 2\noverall score: 1\n\nFor this item, the text prompt is a pirate resembling Blackbeard, and observing an ancient map of the world  ,\nall the frames of video are as follows: \n\n<image> <image> <image> <image> <image> <image> <image> <image> \n"
+            },
+            {
+                "from": "gpt",
+                "value": "visual quality: 1\n object consistency: 1\n dynamic degree: 3\n motion smoothness: 1\n text-to-video alignment: 1\n factual consistency: 1\n overall score: 1\n"
+            }
+        ]
+    },
+    {
+        "id": "a701496",
+        "images": [
+            "a701496_00.jpg",
+            "a701496_01.jpg",
+            "a701496_02.jpg",
+            "a701496_03.jpg",
+            "a701496_04.jpg",
+            "a701496_05.jpg",
+            "a701496_06.jpg",
+            "a701496_07.jpg"
+        ],
+        "conversations": [
+            {
+                "from": "human",
+                "value": "\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos, \nplease watch the following frames of a given video and see the text prompt for generating the video, \nthen give scores from 7 different dimensions:\n(1) visual quality, \n(2) object consistency,\n(3) dynamic degree,\n(4) motion smoothness,\n(5) text-to-video alignment,\n(6) factual consistency, \n(7) overall score\nfor each dimension, output a number from [1,2,3], in which '1' stands for 'Bad', '2' stands for 'Average', '3' stands for 'Good'.\nHere is an output example: \nvisual quality: 3\nobject consistency: 2 \ndynamic degree: 2\nmotion smoothness: 1\ntext-to-video alignment: 1\nfactual consistency: 2\noverall score: 1\n\nFor this item, the text prompt is a painter with a wide brush brushes indigo fabric dye onto wet silk, causing the color to bleed and expand across the canvas  ,\nall the frames of video are as follows: \n\n<image> <image> <image> <image> <image> <image> <image> <image> \n"
+            },
+            {
+                "from": "gpt",
+                "value": "visual quality: 1\n object consistency: 1\n dynamic degree: 3\n motion smoothness: 1\n text-to-video alignment: 1\n factual consistency: 1\n overall score: 1\n"
+            }
+        ]
+    },
+    {
+        "id": "a500560",
+        "images": [
+            "a500560_00.jpg",
+            "a500560_01.jpg",
+            "a500560_02.jpg",
+            "a500560_03.jpg",
+            "a500560_04.jpg",
+            "a500560_05.jpg",
+            "a500560_06.jpg",
+            "a500560_07.jpg",
+            "a500560_08.jpg",
+            "a500560_09.jpg",
+            "a500560_10.jpg",
+            "a500560_11.jpg",
+            "a500560_12.jpg",
+            "a500560_13.jpg",
+            "a500560_14.jpg",
+            "a500560_15.jpg"
+        ],
+        "conversations": [
+            {
+                "from": "human",
+                "value": "\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos, \nplease watch the following frames of a given video and see the text prompt for generating the video, \nthen give scores from 7 different dimensions:\n(1) visual quality, \n(2) object consistency,\n(3) dynamic degree,\n(4) motion smoothness,\n(5) text-to-video alignment,\n(6) factual consistency, \n(7) overall score\nfor each dimension, output a number from [1,2,3], in which '1' stands for 'Bad', '2' stands for 'Average', '3' stands for 'Good'.\nHere is an output example: \nvisual quality: 3\nobject consistency: 2 \ndynamic degree: 2\nmotion smoothness: 1\ntext-to-video alignment: 1\nfactual consistency: 2\noverall score: 1\n\nFor this item, the text prompt is a girl is knitting a doll that looks like a music note, steadly sitting on a cozy chair  ,\nall the frames of video are as follows: \n\n<image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> \n"
+            },
+            {
+                "from": "gpt",
+                "value": "visual quality: 1\n object consistency: 1\n dynamic degree: 3\n motion smoothness: 1\n text-to-video alignment: 1\n factual consistency: 1\n overall score: 1\n"
+            }
+        ]
+    },
+    {
+        "id": "a400721",
+        "images": [
+            "a400721_00.jpg",
+            "a400721_01.jpg",
+            "a400721_02.jpg",
+            "a400721_03.jpg",
+            "a400721_04.jpg",
+            "a400721_05.jpg",
+            "a400721_06.jpg",
+            "a400721_07.jpg",
+            "a400721_08.jpg",
+            "a400721_09.jpg",
+            "a400721_10.jpg",
+            "a400721_11.jpg",
+            "a400721_12.jpg",
+            "a400721_13.jpg",
+            "a400721_14.jpg",
+            "a400721_15.jpg"
+        ],
+        "conversations": [
+            {
+                "from": "human",
+                "value": "\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos, \nplease watch the following frames of a given video and see the text prompt for generating the video, \nthen give scores from 7 different dimensions:\n(1) visual quality, \n(2) object consistency,\n(3) dynamic degree,\n(4) motion smoothness,\n(5) text-to-video alignment,\n(6) factual consistency, \n(7) overall score\nfor each dimension, output a number from [1,2,3], in which '1' stands for 'Bad', '2' stands for 'Average', '3' stands for 'Good'.\nHere is an output example: \nvisual quality: 3\nobject consistency: 2 \ndynamic degree: 2\nmotion smoothness: 1\ntext-to-video alignment: 1\nfactual consistency: 2\noverall score: 1\n\nFor this item, the text prompt is Night sky, black sky, dark blue, slightly purple, bright star in the center, Disney 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> \n"
+            },
+            {
+                "from": "gpt",
+                "value": "visual quality: 1\n object consistency: 1\n dynamic degree: 3\n motion smoothness: 1\n text-to-video alignment: 1\n factual consistency: 1\n overall score: 1\n"
+            }
+        ]
+    },
+    {
+        "id": "a400837",
+        "images": [
+            "a400837_00.jpg",
+            "a400837_01.jpg",
+            "a400837_02.jpg",
+            "a400837_03.jpg",
+            "a400837_04.jpg",
+            "a400837_05.jpg",
+            "a400837_06.jpg",
+            "a400837_07.jpg",
+            "a400837_08.jpg",
+            "a400837_09.jpg",
+            "a400837_10.jpg",
+            "a400837_11.jpg",
+            "a400837_12.jpg",
+            "a400837_13.jpg",
+            "a400837_14.jpg",
+            "a400837_15.jpg"
+        ],
+        "conversations": [
+            {
+                "from": "human",
+                "value": "\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos, \nplease watch the following frames of a given video and see the text prompt for generating the video, \nthen give scores from 7 different dimensions:\n(1) visual quality, \n(2) object consistency,\n(3) dynamic degree,\n(4) motion smoothness,\n(5) text-to-video alignment,\n(6) factual consistency, \n(7) overall score\nfor each dimension, output a number from [1,2,3], in which '1' stands for 'Bad', '2' stands for 'Average', '3' stands for 'Good'.\nHere is an output example: \nvisual quality: 3\nobject consistency: 2 \ndynamic degree: 2\nmotion smoothness: 1\ntext-to-video alignment: 1\nfactual consistency: 2\noverall score: 1\n\nFor this item, the text prompt is carthage ,atlantis civilisation,genral hannibal barca, elephant, port of carthage,army,gold,diamont  ,\nall the frames of video are as follows: \n\n<image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> <image> \n"
+            },
+            {
+                "from": "gpt",
+                "value": "visual quality: 1\n object consistency: 1\n dynamic degree: 3\n motion smoothness: 1\n text-to-video alignment: 1\n factual consistency: 1\n overall score: 1\n"
+            }
+        ]
+    }
+]