Datasets:

CodeGoat24
/

VideoDPO

@article{UnifiedReward,
  title={Unified Reward Model for Multimodal Understanding and Generation.},
  author={Wang, Yibin and Zang, Yuhang, and Li, Hao and Jin, Cheng and Wang Jiaqi},
  journal={arXiv preprint arXiv:2503.05236},
  year={2025}
}

Downloads last month: 68

Size of the auto-converted Parquet files:

3.64 MB

Number of rows:

10,000

Models trained or fine-tuned on CodeGoat24/VideoDPO

CodeGoat24/UnifiedReward-7b

Updated 2 days ago • 98 • 1

Collection including CodeGoat24/VideoDPO

UnifiedReward Training Data

Collection

8 items • Updated 12 days ago • 1

prompt stringlengths 4 8.37k	conversations listlengths 2 2	images sequencelengths 16 16
A nimble anime cat with gleaming eyes, Darts and leaps with playful sighs. Amidst a room of soft, glowing moonlight, Its whimsical dance is a charming sight.	[ { "from": "human", "value": "<image>\n<image>\n<image>\n<image>\n<image>\n<image>\n<image>\n<image>\n <image>\n<image>\n<image>\n<image>\n<image>\n<image>\n<image>\n<image>\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos. You are given a text caption and the frames of two generated videos based on that caption. Your task is to evaluate and compare two videos based on two key criteria:\n1. Alignment with the Caption: Assess how well each video aligns with the provided caption. Consider the accuracy of depicted objects, their relationships, and attributes as described in the caption.\n2. Overall Video Quality: Examine the visual quality of each video, including clarity, detail preservation, color accuracy, and overall aesthetic appeal.\nCompare both videos using the above criteria and select the one that better aligns with the caption while exhibiting superior visual quality.\nProvide a clear conclusion such as \"Video 1 is better than Video 2.\", \"Video 2 is better than Video 1.\" and \"Both videos are equally good.\"\nYour task is provided as follows:\nText Caption: [A nimble anime cat with gleaming eyes,\n Darts and leaps with playful sighs.\n Amidst a room of soft, glowing moonlight,\n Its whimsical dance is a charming sight.]" }, { "from": "gpt", "value": "Video 1 is better than Video 2." } ]	[ "winvideos/006545/0.png", "winvideos/006545/1.png", "winvideos/006545/2.png", "winvideos/006545/3.png", "winvideos/006545/4.png", "winvideos/006545/5.png", "winvideos/006545/6.png", "winvideos/006545/7.png", "losevideos/006545/0.png", "losevideos/006545/1.png", "losevideos/006545/2.png", "losevideos/006545/3.png", "losevideos/006545/4.png", "losevideos/006545/5.png", "losevideos/006545/6.png", "losevideos/006545/7.png" ]
An artist drawing the painting.	[ { "from": "human", "value": "<image>\n<image>\n<image>\n<image>\n<image>\n<image>\n<image>\n<image>\n <image>\n<image>\n<image>\n<image>\n<image>\n<image>\n<image>\n<image>\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos. You are given a text caption and the frames of two generated videos based on that caption. Your task is to evaluate and compare two videos based on two key criteria:\n1. Alignment with the Caption: Assess how well each video aligns with the provided caption. Consider the accuracy of depicted objects, their relationships, and attributes as described in the caption.\n2. Overall Video Quality: Examine the visual quality of each video, including clarity, detail preservation, color accuracy, and overall aesthetic appeal.\nCompare both videos using the above criteria and select the one that better aligns with the caption while exhibiting superior visual quality.\nProvide a clear conclusion such as \"Video 1 is better than Video 2.\", \"Video 2 is better than Video 1.\" and \"Both videos are equally good.\"\nYour task is provided as follows:\nText Caption: [An artist drawing the painting.]" }, { "from": "gpt", "value": "Video 2 is better than Video 1." } ]	[ "losevideos/008145/0.png", "losevideos/008145/1.png", "losevideos/008145/2.png", "losevideos/008145/3.png", "losevideos/008145/4.png", "losevideos/008145/5.png", "losevideos/008145/6.png", "losevideos/008145/7.png", "winvideos/008145/0.png", "winvideos/008145/1.png", "winvideos/008145/2.png", "winvideos/008145/3.png", "winvideos/008145/4.png", "winvideos/008145/5.png", "winvideos/008145/6.png", "winvideos/008145/7.png" ]
bubble waffle with chocolat ice cream.	[ { "from": "human", "value": "<image>\n<image>\n<image>\n<image>\n<image>\n<image>\n<image>\n<image>\n <image>\n<image>\n<image>\n<image>\n<image>\n<image>\n<image>\n<image>\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos. You are given a text caption and the frames of two generated videos based on that caption. Your task is to evaluate and compare two videos based on two key criteria:\n1. Alignment with the Caption: Assess how well each video aligns with the provided caption. Consider the accuracy of depicted objects, their relationships, and attributes as described in the caption.\n2. Overall Video Quality: Examine the visual quality of each video, including clarity, detail preservation, color accuracy, and overall aesthetic appeal.\nCompare both videos using the above criteria and select the one that better aligns with the caption while exhibiting superior visual quality.\nProvide a clear conclusion such as \"Video 1 is better than Video 2.\", \"Video 2 is better than Video 1.\" and \"Both videos are equally good.\"\nYour task is provided as follows:\nText Caption: [bubble waffle with chocolat ice cream.]" }, { "from": "gpt", "value": "Video 1 is better than Video 2." } ]	[ "winvideos/003780/0.png", "winvideos/003780/1.png", "winvideos/003780/2.png", "winvideos/003780/3.png", "winvideos/003780/4.png", "winvideos/003780/5.png", "winvideos/003780/6.png", "winvideos/003780/7.png", "losevideos/003780/0.png", "losevideos/003780/1.png", "losevideos/003780/2.png", "losevideos/003780/3.png", "losevideos/003780/4.png", "losevideos/003780/5.png", "losevideos/003780/6.png", "losevideos/003780/7.png" ]
terrifying floating head approaches quickly.	[ { "from": "human", "value": "<image>\n<image>\n<image>\n<image>\n<image>\n<image>\n<image>\n<image>\n <image>\n<image>\n<image>\n<image>\n<image>\n<image>\n<image>\n<image>\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos. You are given a text caption and the frames of two generated videos based on that caption. Your task is to evaluate and compare two videos based on two key criteria:\n1. Alignment with the Caption: Assess how well each video aligns with the provided caption. Consider the accuracy of depicted objects, their relationships, and attributes as described in the caption.\n2. Overall Video Quality: Examine the visual quality of each video, including clarity, detail preservation, color accuracy, and overall aesthetic appeal.\nCompare both videos using the above criteria and select the one that better aligns with the caption while exhibiting superior visual quality.\nProvide a clear conclusion such as \"Video 1 is better than Video 2.\", \"Video 2 is better than Video 1.\" and \"Both videos are equally good.\"\nYour task is provided as follows:\nText Caption: [terrifying floating head approaches quickly.]" }, { "from": "gpt", "value": "Video 2 is better than Video 1." } ]	[ "losevideos/002941/0.png", "losevideos/002941/1.png", "losevideos/002941/2.png", "losevideos/002941/3.png", "losevideos/002941/4.png", "losevideos/002941/5.png", "losevideos/002941/6.png", "losevideos/002941/7.png", "winvideos/002941/0.png", "winvideos/002941/1.png", "winvideos/002941/2.png", "winvideos/002941/3.png", "winvideos/002941/4.png", "winvideos/002941/5.png", "winvideos/002941/6.png", "winvideos/002941/7.png" ]
The roasted sweet potato is charred and covered in grass ash on top.	[ { "from": "human", "value": "<image>\n<image>\n<image>\n<image>\n<image>\n<image>\n<image>\n<image>\n <image>\n<image>\n<image>\n<image>\n<image>\n<image>\n<image>\n<image>\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos. You are given a text caption and the frames of two generated videos based on that caption. Your task is to evaluate and compare two videos based on two key criteria:\n1. Alignment with the Caption: Assess how well each video aligns with the provided caption. Consider the accuracy of depicted objects, their relationships, and attributes as described in the caption.\n2. Overall Video Quality: Examine the visual quality of each video, including clarity, detail preservation, color accuracy, and overall aesthetic appeal.\nCompare both videos using the above criteria and select the one that better aligns with the caption while exhibiting superior visual quality.\nProvide a clear conclusion such as \"Video 1 is better than Video 2.\", \"Video 2 is better than Video 1.\" and \"Both videos are equally good.\"\nYour task is provided as follows:\nText Caption: [The roasted sweet potato is charred and covered in grass ash on top.]" }, { "from": "gpt", "value": "Video 2 is better than Video 1." } ]	[ "losevideos/002550/0.png", "losevideos/002550/1.png", "losevideos/002550/2.png", "losevideos/002550/3.png", "losevideos/002550/4.png", "losevideos/002550/5.png", "losevideos/002550/6.png", "losevideos/002550/7.png", "winvideos/002550/0.png", "winvideos/002550/1.png", "winvideos/002550/2.png", "winvideos/002550/3.png", "winvideos/002550/4.png", "winvideos/002550/5.png", "winvideos/002550/6.png", "winvideos/002550/7.png" ]
Imagine a tranquil garden setting bathed in the soft glow of morning sunlight. Centered in this image is a vibrant green leaf, almost as big as a hand. On this leaf, a voracious caterpillar is feasting, its tiny mouth parts munching away with determination. The leaf shows signs of being nibbled upon, with small bite marks along its edges. The background should be blurred slightly, putting the focus on the caterpillar and its leafy banquet. One can still make out the shapes of other plants, flowers, and perhaps a dewdrop or two, capturing the essence of a serene morning in nature.".	[ { "from": "human", "value": "<image>\n<image>\n<image>\n<image>\n<image>\n<image>\n<image>\n<image>\n <image>\n<image>\n<image>\n<image>\n<image>\n<image>\n<image>\n<image>\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos. You are given a text caption and the frames of two generated videos based on that caption. Your task is to evaluate and compare two videos based on two key criteria:\n1. Alignment with the Caption: Assess how well each video aligns with the provided caption. Consider the accuracy of depicted objects, their relationships, and attributes as described in the caption.\n2. Overall Video Quality: Examine the visual quality of each video, including clarity, detail preservation, color accuracy, and overall aesthetic appeal.\nCompare both videos using the above criteria and select the one that better aligns with the caption while exhibiting superior visual quality.\nProvide a clear conclusion such as \"Video 1 is better than Video 2.\", \"Video 2 is better than Video 1.\" and \"Both videos are equally good.\"\nYour task is provided as follows:\nText Caption: [Imagine a tranquil garden setting bathed in the soft glow of morning sunlight. Centered in this image is a vibrant green leaf, almost as big as a hand. On this leaf, a voracious caterpillar is feasting, its tiny mouth parts munching away with determination. The leaf shows signs of being nibbled upon, with small bite marks along its edges. The background should be blurred slightly, putting the focus on the caterpillar and its leafy banquet. One can still make out the shapes of other plants, flowers, and perhaps a dewdrop or two, capturing the essence of a serene morning in nature.".]" }, { "from": "gpt", "value": "Video 2 is better than Video 1." } ]	[ "losevideos/008191/0.png", "losevideos/008191/1.png", "losevideos/008191/2.png", "losevideos/008191/3.png", "losevideos/008191/4.png", "losevideos/008191/5.png", "losevideos/008191/6.png", "losevideos/008191/7.png", "winvideos/008191/0.png", "winvideos/008191/1.png", "winvideos/008191/2.png", "winvideos/008191/3.png", "winvideos/008191/4.png", "winvideos/008191/5.png", "winvideos/008191/6.png", "winvideos/008191/7.png" ]
sprinting high.	[ { "from": "human", "value": "<image>\n<image>\n<image>\n<image>\n<image>\n<image>\n<image>\n<image>\n <image>\n<image>\n<image>\n<image>\n<image>\n<image>\n<image>\n<image>\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos. You are given a text caption and the frames of two generated videos based on that caption. Your task is to evaluate and compare two videos based on two key criteria:\n1. Alignment with the Caption: Assess how well each video aligns with the provided caption. Consider the accuracy of depicted objects, their relationships, and attributes as described in the caption.\n2. Overall Video Quality: Examine the visual quality of each video, including clarity, detail preservation, color accuracy, and overall aesthetic appeal.\nCompare both videos using the above criteria and select the one that better aligns with the caption while exhibiting superior visual quality.\nProvide a clear conclusion such as \"Video 1 is better than Video 2.\", \"Video 2 is better than Video 1.\" and \"Both videos are equally good.\"\nYour task is provided as follows:\nText Caption: [sprinting high.]" }, { "from": "gpt", "value": "Video 1 is better than Video 2." } ]	[ "winvideos/006999/0.png", "winvideos/006999/1.png", "winvideos/006999/2.png", "winvideos/006999/3.png", "winvideos/006999/4.png", "winvideos/006999/5.png", "winvideos/006999/6.png", "winvideos/006999/7.png", "losevideos/006999/0.png", "losevideos/006999/1.png", "losevideos/006999/2.png", "losevideos/006999/3.png", "losevideos/006999/4.png", "losevideos/006999/5.png", "losevideos/006999/6.png", "losevideos/006999/7.png" ]
an islamic teacher guy teaching his students in front of the class at the morning in the elementary school.	[ { "from": "human", "value": "<image>\n<image>\n<image>\n<image>\n<image>\n<image>\n<image>\n<image>\n <image>\n<image>\n<image>\n<image>\n<image>\n<image>\n<image>\n<image>\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos. You are given a text caption and the frames of two generated videos based on that caption. Your task is to evaluate and compare two videos based on two key criteria:\n1. Alignment with the Caption: Assess how well each video aligns with the provided caption. Consider the accuracy of depicted objects, their relationships, and attributes as described in the caption.\n2. Overall Video Quality: Examine the visual quality of each video, including clarity, detail preservation, color accuracy, and overall aesthetic appeal.\nCompare both videos using the above criteria and select the one that better aligns with the caption while exhibiting superior visual quality.\nProvide a clear conclusion such as \"Video 1 is better than Video 2.\", \"Video 2 is better than Video 1.\" and \"Both videos are equally good.\"\nYour task is provided as follows:\nText Caption: [an islamic teacher guy teaching his students in front of the class at the morning in the elementary school.]" }, { "from": "gpt", "value": "Video 2 is better than Video 1." } ]	[ "losevideos/008302/0.png", "losevideos/008302/1.png", "losevideos/008302/2.png", "losevideos/008302/3.png", "losevideos/008302/4.png", "losevideos/008302/5.png", "losevideos/008302/6.png", "losevideos/008302/7.png", "winvideos/008302/0.png", "winvideos/008302/1.png", "winvideos/008302/2.png", "winvideos/008302/3.png", "winvideos/008302/4.png", "winvideos/008302/5.png", "winvideos/008302/6.png", "winvideos/008302/7.png" ]
a close up of a person wearing a smiley mask sitting in a chair in a creepy bunker with cameras old crt tv behind, with an upright position looking directly into the camera, movie trailer, female in office dress, evil standing smiling pose, beautiful and realistic faces, , masked, vhs style.	[ { "from": "human", "value": "<image>\n<image>\n<image>\n<image>\n<image>\n<image>\n<image>\n<image>\n <image>\n<image>\n<image>\n<image>\n<image>\n<image>\n<image>\n<image>\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos. You are given a text caption and the frames of two generated videos based on that caption. Your task is to evaluate and compare two videos based on two key criteria:\n1. Alignment with the Caption: Assess how well each video aligns with the provided caption. Consider the accuracy of depicted objects, their relationships, and attributes as described in the caption.\n2. Overall Video Quality: Examine the visual quality of each video, including clarity, detail preservation, color accuracy, and overall aesthetic appeal.\nCompare both videos using the above criteria and select the one that better aligns with the caption while exhibiting superior visual quality.\nProvide a clear conclusion such as \"Video 1 is better than Video 2.\", \"Video 2 is better than Video 1.\" and \"Both videos are equally good.\"\nYour task is provided as follows:\nText Caption: [a close up of a person wearing a smiley mask sitting in a chair in a creepy bunker with cameras old crt tv behind, with an upright position looking directly into the camera, movie trailer, female in office dress, evil standing smiling pose, beautiful and realistic faces, , masked, vhs style.]" }, { "from": "gpt", "value": "Video 1 is better than Video 2." } ]	[ "winvideos/005361/0.png", "winvideos/005361/1.png", "winvideos/005361/2.png", "winvideos/005361/3.png", "winvideos/005361/4.png", "winvideos/005361/5.png", "winvideos/005361/6.png", "winvideos/005361/7.png", "losevideos/005361/0.png", "losevideos/005361/1.png", "losevideos/005361/2.png", "losevideos/005361/3.png", "losevideos/005361/4.png", "losevideos/005361/5.png", "losevideos/005361/6.png", "losevideos/005361/7.png" ]
explosion, black background -motion 3.	[ { "from": "human", "value": "<image>\n<image>\n<image>\n<image>\n<image>\n<image>\n<image>\n<image>\n <image>\n<image>\n<image>\n<image>\n<image>\n<image>\n<image>\n<image>\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos. You are given a text caption and the frames of two generated videos based on that caption. Your task is to evaluate and compare two videos based on two key criteria:\n1. Alignment with the Caption: Assess how well each video aligns with the provided caption. Consider the accuracy of depicted objects, their relationships, and attributes as described in the caption.\n2. Overall Video Quality: Examine the visual quality of each video, including clarity, detail preservation, color accuracy, and overall aesthetic appeal.\nCompare both videos using the above criteria and select the one that better aligns with the caption while exhibiting superior visual quality.\nProvide a clear conclusion such as \"Video 1 is better than Video 2.\", \"Video 2 is better than Video 1.\" and \"Both videos are equally good.\"\nYour task is provided as follows:\nText Caption: [explosion, black background -motion 3.]" }, { "from": "gpt", "value": "Video 1 is better than Video 2." } ]	[ "winvideos/000925/0.png", "winvideos/000925/1.png", "winvideos/000925/2.png", "winvideos/000925/3.png", "winvideos/000925/4.png", "winvideos/000925/5.png", "winvideos/000925/6.png", "winvideos/000925/7.png", "losevideos/000925/0.png", "losevideos/000925/1.png", "losevideos/000925/2.png", "losevideos/000925/3.png", "losevideos/000925/4.png", "losevideos/000925/5.png", "losevideos/000925/6.png", "losevideos/000925/7.png" ]
the man sneezed sharply and flew away.	[ { "from": "human", "value": "<image>\n<image>\n<image>\n<image>\n<image>\n<image>\n<image>\n<image>\n <image>\n<image>\n<image>\n<image>\n<image>\n<image>\n<image>\n<image>\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos. You are given a text caption and the frames of two generated videos based on that caption. Your task is to evaluate and compare two videos based on two key criteria:\n1. Alignment with the Caption: Assess how well each video aligns with the provided caption. Consider the accuracy of depicted objects, their relationships, and attributes as described in the caption.\n2. Overall Video Quality: Examine the visual quality of each video, including clarity, detail preservation, color accuracy, and overall aesthetic appeal.\nCompare both videos using the above criteria and select the one that better aligns with the caption while exhibiting superior visual quality.\nProvide a clear conclusion such as \"Video 1 is better than Video 2.\", \"Video 2 is better than Video 1.\" and \"Both videos are equally good.\"\nYour task is provided as follows:\nText Caption: [the man sneezed sharply and flew away.]" }, { "from": "gpt", "value": "Video 1 is better than Video 2." } ]	[ "winvideos/000863/0.png", "winvideos/000863/1.png", "winvideos/000863/2.png", "winvideos/000863/3.png", "winvideos/000863/4.png", "winvideos/000863/5.png", "winvideos/000863/6.png", "winvideos/000863/7.png", "losevideos/000863/0.png", "losevideos/000863/1.png", "losevideos/000863/2.png", "losevideos/000863/3.png", "losevideos/000863/4.png", "losevideos/000863/5.png", "losevideos/000863/6.png", "losevideos/000863/7.png" ]
Create a photorealistic 8k portrait photo of the most beautiful 20 years old hybrid 50% fairy and 50% woman, zoom 30 mm, teleobjectif of far character aimed, long shot, from feet view, whole character view, light from behind, translucency, -neg "worst resolution, ugly face, extra-limbs".	[ { "from": "human", "value": "<image>\n<image>\n<image>\n<image>\n<image>\n<image>\n<image>\n<image>\n <image>\n<image>\n<image>\n<image>\n<image>\n<image>\n<image>\n<image>\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos. You are given a text caption and the frames of two generated videos based on that caption. Your task is to evaluate and compare two videos based on two key criteria:\n1. Alignment with the Caption: Assess how well each video aligns with the provided caption. Consider the accuracy of depicted objects, their relationships, and attributes as described in the caption.\n2. Overall Video Quality: Examine the visual quality of each video, including clarity, detail preservation, color accuracy, and overall aesthetic appeal.\nCompare both videos using the above criteria and select the one that better aligns with the caption while exhibiting superior visual quality.\nProvide a clear conclusion such as \"Video 1 is better than Video 2.\", \"Video 2 is better than Video 1.\" and \"Both videos are equally good.\"\nYour task is provided as follows:\nText Caption: [Create a photorealistic 8k portrait photo of the most beautiful 20 years old hybrid 50% fairy and 50% woman, zoom 30 mm, teleobjectif of far character aimed, long shot, from feet view, whole character view, light from behind, translucency, -neg "worst resolution, ugly face, extra-limbs".]" }, { "from": "gpt", "value": "Video 2 is better than Video 1." } ]	[ "losevideos/001642/0.png", "losevideos/001642/1.png", "losevideos/001642/2.png", "losevideos/001642/3.png", "losevideos/001642/4.png", "losevideos/001642/5.png", "losevideos/001642/6.png", "losevideos/001642/7.png", "winvideos/001642/0.png", "winvideos/001642/1.png", "winvideos/001642/2.png", "winvideos/001642/3.png", "winvideos/001642/4.png", "winvideos/001642/5.png", "winvideos/001642/6.png", "winvideos/001642/7.png" ]
The protagonist faces a formidable, menacing foe. A thrilling battle or confrontation filled with suspense.	[ { "from": "human", "value": "<image>\n<image>\n<image>\n<image>\n<image>\n<image>\n<image>\n<image>\n <image>\n<image>\n<image>\n<image>\n<image>\n<image>\n<image>\n<image>\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos. You are given a text caption and the frames of two generated videos based on that caption. Your task is to evaluate and compare two videos based on two key criteria:\n1. Alignment with the Caption: Assess how well each video aligns with the provided caption. Consider the accuracy of depicted objects, their relationships, and attributes as described in the caption.\n2. Overall Video Quality: Examine the visual quality of each video, including clarity, detail preservation, color accuracy, and overall aesthetic appeal.\nCompare both videos using the above criteria and select the one that better aligns with the caption while exhibiting superior visual quality.\nProvide a clear conclusion such as \"Video 1 is better than Video 2.\", \"Video 2 is better than Video 1.\" and \"Both videos are equally good.\"\nYour task is provided as follows:\nText Caption: [The protagonist faces a formidable, menacing foe.\nA thrilling battle or confrontation filled with suspense.]" }, { "from": "gpt", "value": "Video 1 is better than Video 2." } ]	[ "winvideos/007441/0.png", "winvideos/007441/1.png", "winvideos/007441/2.png", "winvideos/007441/3.png", "winvideos/007441/4.png", "winvideos/007441/5.png", "winvideos/007441/6.png", "winvideos/007441/7.png", "losevideos/007441/0.png", "losevideos/007441/1.png", "losevideos/007441/2.png", "losevideos/007441/3.png", "losevideos/007441/4.png", "losevideos/007441/5.png", "losevideos/007441/6.png", "losevideos/007441/7.png" ]
Moreover, the discovery of ancient humans.	[ { "from": "human", "value": "<image>\n<image>\n<image>\n<image>\n<image>\n<image>\n<image>\n<image>\n <image>\n<image>\n<image>\n<image>\n<image>\n<image>\n<image>\n<image>\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos. You are given a text caption and the frames of two generated videos based on that caption. Your task is to evaluate and compare two videos based on two key criteria:\n1. Alignment with the Caption: Assess how well each video aligns with the provided caption. Consider the accuracy of depicted objects, their relationships, and attributes as described in the caption.\n2. Overall Video Quality: Examine the visual quality of each video, including clarity, detail preservation, color accuracy, and overall aesthetic appeal.\nCompare both videos using the above criteria and select the one that better aligns with the caption while exhibiting superior visual quality.\nProvide a clear conclusion such as \"Video 1 is better than Video 2.\", \"Video 2 is better than Video 1.\" and \"Both videos are equally good.\"\nYour task is provided as follows:\nText Caption: [Moreover, the discovery of ancient humans.]" }, { "from": "gpt", "value": "Video 2 is better than Video 1." } ]	[ "losevideos/005731/0.png", "losevideos/005731/1.png", "losevideos/005731/2.png", "losevideos/005731/3.png", "losevideos/005731/4.png", "losevideos/005731/5.png", "losevideos/005731/6.png", "losevideos/005731/7.png", "winvideos/005731/0.png", "winvideos/005731/1.png", "winvideos/005731/2.png", "winvideos/005731/3.png", "winvideos/005731/4.png", "winvideos/005731/5.png", "winvideos/005731/6.png", "winvideos/005731/7.png" ]
A girl with long light blue hair and bright blue eyes, lightning and fire Powers, moving, animation, cinematic -gs 20 -motion 4 image -realism.	[ { "from": "human", "value": "<image>\n<image>\n<image>\n<image>\n<image>\n<image>\n<image>\n<image>\n <image>\n<image>\n<image>\n<image>\n<image>\n<image>\n<image>\n<image>\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos. You are given a text caption and the frames of two generated videos based on that caption. Your task is to evaluate and compare two videos based on two key criteria:\n1. Alignment with the Caption: Assess how well each video aligns with the provided caption. Consider the accuracy of depicted objects, their relationships, and attributes as described in the caption.\n2. Overall Video Quality: Examine the visual quality of each video, including clarity, detail preservation, color accuracy, and overall aesthetic appeal.\nCompare both videos using the above criteria and select the one that better aligns with the caption while exhibiting superior visual quality.\nProvide a clear conclusion such as \"Video 1 is better than Video 2.\", \"Video 2 is better than Video 1.\" and \"Both videos are equally good.\"\nYour task is provided as follows:\nText Caption: [A girl with long light blue hair and bright blue eyes, lightning and fire Powers, moving, animation, cinematic -gs 20 -motion 4 image -realism.]" }, { "from": "gpt", "value": "Video 1 is better than Video 2." } ]	[ "winvideos/003628/0.png", "winvideos/003628/1.png", "winvideos/003628/2.png", "winvideos/003628/3.png", "winvideos/003628/4.png", "winvideos/003628/5.png", "winvideos/003628/6.png", "winvideos/003628/7.png", "losevideos/003628/0.png", "losevideos/003628/1.png", "losevideos/003628/2.png", "losevideos/003628/3.png", "losevideos/003628/4.png", "losevideos/003628/5.png", "losevideos/003628/6.png", "losevideos/003628/7.png" ]
atardecer en la playa, fieste elegante, caminando toda la orilla hasta llegar al final del horizonte.	[ { "from": "human", "value": "<image>\n<image>\n<image>\n<image>\n<image>\n<image>\n<image>\n<image>\n <image>\n<image>\n<image>\n<image>\n<image>\n<image>\n<image>\n<image>\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos. You are given a text caption and the frames of two generated videos based on that caption. Your task is to evaluate and compare two videos based on two key criteria:\n1. Alignment with the Caption: Assess how well each video aligns with the provided caption. Consider the accuracy of depicted objects, their relationships, and attributes as described in the caption.\n2. Overall Video Quality: Examine the visual quality of each video, including clarity, detail preservation, color accuracy, and overall aesthetic appeal.\nCompare both videos using the above criteria and select the one that better aligns with the caption while exhibiting superior visual quality.\nProvide a clear conclusion such as \"Video 1 is better than Video 2.\", \"Video 2 is better than Video 1.\" and \"Both videos are equally good.\"\nYour task is provided as follows:\nText Caption: [atardecer en la playa, fieste elegante, caminando toda la orilla hasta llegar al final del horizonte.]" }, { "from": "gpt", "value": "Video 2 is better than Video 1." } ]	[ "losevideos/005360/0.png", "losevideos/005360/1.png", "losevideos/005360/2.png", "losevideos/005360/3.png", "losevideos/005360/4.png", "losevideos/005360/5.png", "losevideos/005360/6.png", "losevideos/005360/7.png", "winvideos/005360/0.png", "winvideos/005360/1.png", "winvideos/005360/2.png", "winvideos/005360/3.png", "winvideos/005360/4.png", "winvideos/005360/5.png", "winvideos/005360/6.png", "winvideos/005360/7.png" ]
Ancient 1504 AD era, A Swahili woman and a Portuguese man sit down to break bread together, Super Detail, Ultra HD, 16k -fps 24 -gs 18 -motion 3 -neg deformed, bad anatomy, disfigured, poorly drawn face, mutation, mutated, extra limb, ugly, disgusting, poorly drawn hands, missing limb, floating limbs, disconnected limbs, malformed hands, blurry, morphing, erratic fluctuation in motion, noisy, bad quality, distorted, poorly drawn, blurry, grainy, low resolution, over saturated, lack of detail, inconsistent lighting, out of frame, cropped bad proportion.	[ { "from": "human", "value": "<image>\n<image>\n<image>\n<image>\n<image>\n<image>\n<image>\n<image>\n <image>\n<image>\n<image>\n<image>\n<image>\n<image>\n<image>\n<image>\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos. You are given a text caption and the frames of two generated videos based on that caption. Your task is to evaluate and compare two videos based on two key criteria:\n1. Alignment with the Caption: Assess how well each video aligns with the provided caption. Consider the accuracy of depicted objects, their relationships, and attributes as described in the caption.\n2. Overall Video Quality: Examine the visual quality of each video, including clarity, detail preservation, color accuracy, and overall aesthetic appeal.\nCompare both videos using the above criteria and select the one that better aligns with the caption while exhibiting superior visual quality.\nProvide a clear conclusion such as \"Video 1 is better than Video 2.\", \"Video 2 is better than Video 1.\" and \"Both videos are equally good.\"\nYour task is provided as follows:\nText Caption: [Ancient 1504 AD era, A Swahili woman and a Portuguese man sit down to break bread together, Super Detail, Ultra HD, 16k -fps 24 -gs 18 -motion 3 -neg deformed, bad anatomy, disfigured, poorly drawn face, mutation, mutated, extra limb, ugly, disgusting, poorly drawn hands, missing limb, floating limbs, disconnected limbs, malformed hands, blurry, morphing, erratic fluctuation in motion, noisy, bad quality, distorted, poorly drawn, blurry, grainy, low resolution, over saturated, lack of detail, inconsistent lighting, out of frame, cropped bad proportion.]" }, { "from": "gpt", "value": "Video 2 is better than Video 1." } ]	[ "losevideos/009893/0.png", "losevideos/009893/1.png", "losevideos/009893/2.png", "losevideos/009893/3.png", "losevideos/009893/4.png", "losevideos/009893/5.png", "losevideos/009893/6.png", "losevideos/009893/7.png", "winvideos/009893/0.png", "winvideos/009893/1.png", "winvideos/009893/2.png", "winvideos/009893/3.png", "winvideos/009893/4.png", "winvideos/009893/5.png", "winvideos/009893/6.png", "winvideos/009893/7.png" ]
had moving -motion 2 -ar 9:18.	[ { "from": "human", "value": "<image>\n<image>\n<image>\n<image>\n<image>\n<image>\n<image>\n<image>\n <image>\n<image>\n<image>\n<image>\n<image>\n<image>\n<image>\n<image>\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos. You are given a text caption and the frames of two generated videos based on that caption. Your task is to evaluate and compare two videos based on two key criteria:\n1. Alignment with the Caption: Assess how well each video aligns with the provided caption. Consider the accuracy of depicted objects, their relationships, and attributes as described in the caption.\n2. Overall Video Quality: Examine the visual quality of each video, including clarity, detail preservation, color accuracy, and overall aesthetic appeal.\nCompare both videos using the above criteria and select the one that better aligns with the caption while exhibiting superior visual quality.\nProvide a clear conclusion such as \"Video 1 is better than Video 2.\", \"Video 2 is better than Video 1.\" and \"Both videos are equally good.\"\nYour task is provided as follows:\nText Caption: [had moving -motion 2 -ar 9:18.]" }, { "from": "gpt", "value": "Video 1 is better than Video 2." } ]	[ "winvideos/009508/0.png", "winvideos/009508/1.png", "winvideos/009508/2.png", "winvideos/009508/3.png", "winvideos/009508/4.png", "winvideos/009508/5.png", "winvideos/009508/6.png", "winvideos/009508/7.png", "losevideos/009508/0.png", "losevideos/009508/1.png", "losevideos/009508/2.png", "losevideos/009508/3.png", "losevideos/009508/4.png", "losevideos/009508/5.png", "losevideos/009508/6.png", "losevideos/009508/7.png" ]
Improve this.	[ { "from": "human", "value": "<image>\n<image>\n<image>\n<image>\n<image>\n<image>\n<image>\n<image>\n <image>\n<image>\n<image>\n<image>\n<image>\n<image>\n<image>\n<image>\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos. You are given a text caption and the frames of two generated videos based on that caption. Your task is to evaluate and compare two videos based on two key criteria:\n1. Alignment with the Caption: Assess how well each video aligns with the provided caption. Consider the accuracy of depicted objects, their relationships, and attributes as described in the caption.\n2. Overall Video Quality: Examine the visual quality of each video, including clarity, detail preservation, color accuracy, and overall aesthetic appeal.\nCompare both videos using the above criteria and select the one that better aligns with the caption while exhibiting superior visual quality.\nProvide a clear conclusion such as \"Video 1 is better than Video 2.\", \"Video 2 is better than Video 1.\" and \"Both videos are equally good.\"\nYour task is provided as follows:\nText Caption: [Improve this.]" }, { "from": "gpt", "value": "Video 1 is better than Video 2." } ]	[ "winvideos/008206/0.png", "winvideos/008206/1.png", "winvideos/008206/2.png", "winvideos/008206/3.png", "winvideos/008206/4.png", "winvideos/008206/5.png", "winvideos/008206/6.png", "winvideos/008206/7.png", "losevideos/008206/0.png", "losevideos/008206/1.png", "losevideos/008206/2.png", "losevideos/008206/3.png", "losevideos/008206/4.png", "losevideos/008206/5.png", "losevideos/008206/6.png", "losevideos/008206/7.png" ]
paradise a view from the moon of earth, cinematic opening sequence, establishing shot, empty space with mars in the center, spaceship orbiting mars, moons visible, super resolution -- ar 16:9.	[ { "from": "human", "value": "<image>\n<image>\n<image>\n<image>\n<image>\n<image>\n<image>\n<image>\n <image>\n<image>\n<image>\n<image>\n<image>\n<image>\n<image>\n<image>\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos. You are given a text caption and the frames of two generated videos based on that caption. Your task is to evaluate and compare two videos based on two key criteria:\n1. Alignment with the Caption: Assess how well each video aligns with the provided caption. Consider the accuracy of depicted objects, their relationships, and attributes as described in the caption.\n2. Overall Video Quality: Examine the visual quality of each video, including clarity, detail preservation, color accuracy, and overall aesthetic appeal.\nCompare both videos using the above criteria and select the one that better aligns with the caption while exhibiting superior visual quality.\nProvide a clear conclusion such as \"Video 1 is better than Video 2.\", \"Video 2 is better than Video 1.\" and \"Both videos are equally good.\"\nYour task is provided as follows:\nText Caption: [paradise a view from the moon of earth, cinematic opening sequence, establishing shot, empty space with mars in the center, spaceship orbiting mars, moons visible, super resolution -- ar 16:9.]" }, { "from": "gpt", "value": "Video 2 is better than Video 1." } ]	[ "losevideos/008152/0.png", "losevideos/008152/1.png", "losevideos/008152/2.png", "losevideos/008152/3.png", "losevideos/008152/4.png", "losevideos/008152/5.png", "losevideos/008152/6.png", "losevideos/008152/7.png", "winvideos/008152/0.png", "winvideos/008152/1.png", "winvideos/008152/2.png", "winvideos/008152/3.png", "winvideos/008152/4.png", "winvideos/008152/5.png", "winvideos/008152/6.png", "winvideos/008152/7.png" ]
shoes lying on a stage, comic style, colorful, rock'n'roll, stage lights.	[ { "from": "human", "value": "<image>\n<image>\n<image>\n<image>\n<image>\n<image>\n<image>\n<image>\n <image>\n<image>\n<image>\n<image>\n<image>\n<image>\n<image>\n<image>\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos. You are given a text caption and the frames of two generated videos based on that caption. Your task is to evaluate and compare two videos based on two key criteria:\n1. Alignment with the Caption: Assess how well each video aligns with the provided caption. Consider the accuracy of depicted objects, their relationships, and attributes as described in the caption.\n2. Overall Video Quality: Examine the visual quality of each video, including clarity, detail preservation, color accuracy, and overall aesthetic appeal.\nCompare both videos using the above criteria and select the one that better aligns with the caption while exhibiting superior visual quality.\nProvide a clear conclusion such as \"Video 1 is better than Video 2.\", \"Video 2 is better than Video 1.\" and \"Both videos are equally good.\"\nYour task is provided as follows:\nText Caption: [shoes lying on a stage, comic style, colorful, rock'n'roll, stage lights.]" }, { "from": "gpt", "value": "Video 1 is better than Video 2." } ]	[ "winvideos/001815/0.png", "winvideos/001815/1.png", "winvideos/001815/2.png", "winvideos/001815/3.png", "winvideos/001815/4.png", "winvideos/001815/5.png", "winvideos/001815/6.png", "winvideos/001815/7.png", "losevideos/001815/0.png", "losevideos/001815/1.png", "losevideos/001815/2.png", "losevideos/001815/3.png", "losevideos/001815/4.png", "losevideos/001815/5.png", "losevideos/001815/6.png", "losevideos/001815/7.png" ]
Show bustling city streets filled with people in business attire. The streets should be crowded with individuals in a hurry, exchanging envelopes discreetly. Convey the sense of secrecy and suspicious transactions happening all around.	[ { "from": "human", "value": "<image>\n<image>\n<image>\n<image>\n<image>\n<image>\n<image>\n<image>\n <image>\n<image>\n<image>\n<image>\n<image>\n<image>\n<image>\n<image>\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos. You are given a text caption and the frames of two generated videos based on that caption. Your task is to evaluate and compare two videos based on two key criteria:\n1. Alignment with the Caption: Assess how well each video aligns with the provided caption. Consider the accuracy of depicted objects, their relationships, and attributes as described in the caption.\n2. Overall Video Quality: Examine the visual quality of each video, including clarity, detail preservation, color accuracy, and overall aesthetic appeal.\nCompare both videos using the above criteria and select the one that better aligns with the caption while exhibiting superior visual quality.\nProvide a clear conclusion such as \"Video 1 is better than Video 2.\", \"Video 2 is better than Video 1.\" and \"Both videos are equally good.\"\nYour task is provided as follows:\nText Caption: [Show bustling city streets filled with people in business attire. The streets should be crowded with individuals in a hurry, exchanging envelopes discreetly. Convey the sense of secrecy and suspicious transactions happening all around.]" }, { "from": "gpt", "value": "Video 2 is better than Video 1." } ]	[ "losevideos/006791/0.png", "losevideos/006791/1.png", "losevideos/006791/2.png", "losevideos/006791/3.png", "losevideos/006791/4.png", "losevideos/006791/5.png", "losevideos/006791/6.png", "losevideos/006791/7.png", "winvideos/006791/0.png", "winvideos/006791/1.png", "winvideos/006791/2.png", "winvideos/006791/3.png", "winvideos/006791/4.png", "winvideos/006791/5.png", "winvideos/006791/6.png", "winvideos/006791/7.png" ]
Panda eating an empanada at a moonlit table.	[ { "from": "human", "value": "<image>\n<image>\n<image>\n<image>\n<image>\n<image>\n<image>\n<image>\n <image>\n<image>\n<image>\n<image>\n<image>\n<image>\n<image>\n<image>\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos. You are given a text caption and the frames of two generated videos based on that caption. Your task is to evaluate and compare two videos based on two key criteria:\n1. Alignment with the Caption: Assess how well each video aligns with the provided caption. Consider the accuracy of depicted objects, their relationships, and attributes as described in the caption.\n2. Overall Video Quality: Examine the visual quality of each video, including clarity, detail preservation, color accuracy, and overall aesthetic appeal.\nCompare both videos using the above criteria and select the one that better aligns with the caption while exhibiting superior visual quality.\nProvide a clear conclusion such as \"Video 1 is better than Video 2.\", \"Video 2 is better than Video 1.\" and \"Both videos are equally good.\"\nYour task is provided as follows:\nText Caption: [Panda eating an empanada at a moonlit table.]" }, { "from": "gpt", "value": "Video 2 is better than Video 1." } ]	[ "losevideos/007955/0.png", "losevideos/007955/1.png", "losevideos/007955/2.png", "losevideos/007955/3.png", "losevideos/007955/4.png", "losevideos/007955/5.png", "losevideos/007955/6.png", "losevideos/007955/7.png", "winvideos/007955/0.png", "winvideos/007955/1.png", "winvideos/007955/2.png", "winvideos/007955/3.png", "winvideos/007955/4.png", "winvideos/007955/5.png", "winvideos/007955/6.png", "winvideos/007955/7.png" ]
beautifull shiny fish with other see creature like Disney in 8k.	[ { "from": "human", "value": "<image>\n<image>\n<image>\n<image>\n<image>\n<image>\n<image>\n<image>\n <image>\n<image>\n<image>\n<image>\n<image>\n<image>\n<image>\n<image>\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos. You are given a text caption and the frames of two generated videos based on that caption. Your task is to evaluate and compare two videos based on two key criteria:\n1. Alignment with the Caption: Assess how well each video aligns with the provided caption. Consider the accuracy of depicted objects, their relationships, and attributes as described in the caption.\n2. Overall Video Quality: Examine the visual quality of each video, including clarity, detail preservation, color accuracy, and overall aesthetic appeal.\nCompare both videos using the above criteria and select the one that better aligns with the caption while exhibiting superior visual quality.\nProvide a clear conclusion such as \"Video 1 is better than Video 2.\", \"Video 2 is better than Video 1.\" and \"Both videos are equally good.\"\nYour task is provided as follows:\nText Caption: [beautifull shiny fish with other see creature like Disney in 8k.]" }, { "from": "gpt", "value": "Video 1 is better than Video 2." } ]	[ "winvideos/004173/0.png", "winvideos/004173/1.png", "winvideos/004173/2.png", "winvideos/004173/3.png", "winvideos/004173/4.png", "winvideos/004173/5.png", "winvideos/004173/6.png", "winvideos/004173/7.png", "losevideos/004173/0.png", "losevideos/004173/1.png", "losevideos/004173/2.png", "losevideos/004173/3.png", "losevideos/004173/4.png", "losevideos/004173/5.png", "losevideos/004173/6.png", "losevideos/004173/7.png" ]
a Lancashire Heeler shout at cow.	[ { "from": "human", "value": "<image>\n<image>\n<image>\n<image>\n<image>\n<image>\n<image>\n<image>\n <image>\n<image>\n<image>\n<image>\n<image>\n<image>\n<image>\n<image>\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos. You are given a text caption and the frames of two generated videos based on that caption. Your task is to evaluate and compare two videos based on two key criteria:\n1. Alignment with the Caption: Assess how well each video aligns with the provided caption. Consider the accuracy of depicted objects, their relationships, and attributes as described in the caption.\n2. Overall Video Quality: Examine the visual quality of each video, including clarity, detail preservation, color accuracy, and overall aesthetic appeal.\nCompare both videos using the above criteria and select the one that better aligns with the caption while exhibiting superior visual quality.\nProvide a clear conclusion such as \"Video 1 is better than Video 2.\", \"Video 2 is better than Video 1.\" and \"Both videos are equally good.\"\nYour task is provided as follows:\nText Caption: [a Lancashire Heeler shout at cow.]" }, { "from": "gpt", "value": "Video 2 is better than Video 1." } ]	[ "losevideos/005963/0.png", "losevideos/005963/1.png", "losevideos/005963/2.png", "losevideos/005963/3.png", "losevideos/005963/4.png", "losevideos/005963/5.png", "losevideos/005963/6.png", "losevideos/005963/7.png", "winvideos/005963/0.png", "winvideos/005963/1.png", "winvideos/005963/2.png", "winvideos/005963/3.png", "winvideos/005963/4.png", "winvideos/005963/5.png", "winvideos/005963/6.png", "winvideos/005963/7.png" ]
a growing tree --creative --test --ar 16:9 --video--v3.	[ { "from": "human", "value": "<image>\n<image>\n<image>\n<image>\n<image>\n<image>\n<image>\n<image>\n <image>\n<image>\n<image>\n<image>\n<image>\n<image>\n<image>\n<image>\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos. You are given a text caption and the frames of two generated videos based on that caption. Your task is to evaluate and compare two videos based on two key criteria:\n1. Alignment with the Caption: Assess how well each video aligns with the provided caption. Consider the accuracy of depicted objects, their relationships, and attributes as described in the caption.\n2. Overall Video Quality: Examine the visual quality of each video, including clarity, detail preservation, color accuracy, and overall aesthetic appeal.\nCompare both videos using the above criteria and select the one that better aligns with the caption while exhibiting superior visual quality.\nProvide a clear conclusion such as \"Video 1 is better than Video 2.\", \"Video 2 is better than Video 1.\" and \"Both videos are equally good.\"\nYour task is provided as follows:\nText Caption: [a growing tree --creative --test --ar 16:9 --video--v3.]" }, { "from": "gpt", "value": "Video 1 is better than Video 2." } ]	[ "winvideos/004701/0.png", "winvideos/004701/1.png", "winvideos/004701/2.png", "winvideos/004701/3.png", "winvideos/004701/4.png", "winvideos/004701/5.png", "winvideos/004701/6.png", "winvideos/004701/7.png", "losevideos/004701/0.png", "losevideos/004701/1.png", "losevideos/004701/2.png", "losevideos/004701/3.png", "losevideos/004701/4.png", "losevideos/004701/5.png", "losevideos/004701/6.png", "losevideos/004701/7.png" ]
un grupo de 5 personas realizando actividad fisica pero con el pasar del tiempo van envejeciendo saludablemente con cuerpos tonificados y esbeltos.	[ { "from": "human", "value": "<image>\n<image>\n<image>\n<image>\n<image>\n<image>\n<image>\n<image>\n <image>\n<image>\n<image>\n<image>\n<image>\n<image>\n<image>\n<image>\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos. You are given a text caption and the frames of two generated videos based on that caption. Your task is to evaluate and compare two videos based on two key criteria:\n1. Alignment with the Caption: Assess how well each video aligns with the provided caption. Consider the accuracy of depicted objects, their relationships, and attributes as described in the caption.\n2. Overall Video Quality: Examine the visual quality of each video, including clarity, detail preservation, color accuracy, and overall aesthetic appeal.\nCompare both videos using the above criteria and select the one that better aligns with the caption while exhibiting superior visual quality.\nProvide a clear conclusion such as \"Video 1 is better than Video 2.\", \"Video 2 is better than Video 1.\" and \"Both videos are equally good.\"\nYour task is provided as follows:\nText Caption: [un grupo de 5 personas realizando actividad fisica pero con el pasar del tiempo van envejeciendo saludablemente con cuerpos tonificados y esbeltos.]" }, { "from": "gpt", "value": "Video 2 is better than Video 1." } ]	[ "losevideos/009395/0.png", "losevideos/009395/1.png", "losevideos/009395/2.png", "losevideos/009395/3.png", "losevideos/009395/4.png", "losevideos/009395/5.png", "losevideos/009395/6.png", "losevideos/009395/7.png", "winvideos/009395/0.png", "winvideos/009395/1.png", "winvideos/009395/2.png", "winvideos/009395/3.png", "winvideos/009395/4.png", "winvideos/009395/5.png", "winvideos/009395/6.png", "winvideos/009395/7.png" ]
a demon running through fiery hell.	[ { "from": "human", "value": "<image>\n<image>\n<image>\n<image>\n<image>\n<image>\n<image>\n<image>\n <image>\n<image>\n<image>\n<image>\n<image>\n<image>\n<image>\n<image>\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos. You are given a text caption and the frames of two generated videos based on that caption. Your task is to evaluate and compare two videos based on two key criteria:\n1. Alignment with the Caption: Assess how well each video aligns with the provided caption. Consider the accuracy of depicted objects, their relationships, and attributes as described in the caption.\n2. Overall Video Quality: Examine the visual quality of each video, including clarity, detail preservation, color accuracy, and overall aesthetic appeal.\nCompare both videos using the above criteria and select the one that better aligns with the caption while exhibiting superior visual quality.\nProvide a clear conclusion such as \"Video 1 is better than Video 2.\", \"Video 2 is better than Video 1.\" and \"Both videos are equally good.\"\nYour task is provided as follows:\nText Caption: [a demon running through fiery hell.]" }, { "from": "gpt", "value": "Video 2 is better than Video 1." } ]	[ "losevideos/002223/0.png", "losevideos/002223/1.png", "losevideos/002223/2.png", "losevideos/002223/3.png", "losevideos/002223/4.png", "losevideos/002223/5.png", "losevideos/002223/6.png", "losevideos/002223/7.png", "winvideos/002223/0.png", "winvideos/002223/1.png", "winvideos/002223/2.png", "winvideos/002223/3.png", "winvideos/002223/4.png", "winvideos/002223/5.png", "winvideos/002223/6.png", "winvideos/002223/7.png" ]
Portray the escalating panic as the friends attempt to escape. 3d animation, 16.9 aspect.	[ { "from": "human", "value": "<image>\n<image>\n<image>\n<image>\n<image>\n<image>\n<image>\n<image>\n <image>\n<image>\n<image>\n<image>\n<image>\n<image>\n<image>\n<image>\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos. You are given a text caption and the frames of two generated videos based on that caption. Your task is to evaluate and compare two videos based on two key criteria:\n1. Alignment with the Caption: Assess how well each video aligns with the provided caption. Consider the accuracy of depicted objects, their relationships, and attributes as described in the caption.\n2. Overall Video Quality: Examine the visual quality of each video, including clarity, detail preservation, color accuracy, and overall aesthetic appeal.\nCompare both videos using the above criteria and select the one that better aligns with the caption while exhibiting superior visual quality.\nProvide a clear conclusion such as \"Video 1 is better than Video 2.\", \"Video 2 is better than Video 1.\" and \"Both videos are equally good.\"\nYour task is provided as follows:\nText Caption: [Portray the escalating panic as the friends attempt to escape. 3d animation, 16.9 aspect.]" }, { "from": "gpt", "value": "Video 1 is better than Video 2." } ]	[ "winvideos/009875/0.png", "winvideos/009875/1.png", "winvideos/009875/2.png", "winvideos/009875/3.png", "winvideos/009875/4.png", "winvideos/009875/5.png", "winvideos/009875/6.png", "winvideos/009875/7.png", "losevideos/009875/0.png", "losevideos/009875/1.png", "losevideos/009875/2.png", "losevideos/009875/3.png", "losevideos/009875/4.png", "losevideos/009875/5.png", "losevideos/009875/6.png", "losevideos/009875/7.png" ]
virtual dressing room.	[ { "from": "human", "value": "<image>\n<image>\n<image>\n<image>\n<image>\n<image>\n<image>\n<image>\n <image>\n<image>\n<image>\n<image>\n<image>\n<image>\n<image>\n<image>\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos. You are given a text caption and the frames of two generated videos based on that caption. Your task is to evaluate and compare two videos based on two key criteria:\n1. Alignment with the Caption: Assess how well each video aligns with the provided caption. Consider the accuracy of depicted objects, their relationships, and attributes as described in the caption.\n2. Overall Video Quality: Examine the visual quality of each video, including clarity, detail preservation, color accuracy, and overall aesthetic appeal.\nCompare both videos using the above criteria and select the one that better aligns with the caption while exhibiting superior visual quality.\nProvide a clear conclusion such as \"Video 1 is better than Video 2.\", \"Video 2 is better than Video 1.\" and \"Both videos are equally good.\"\nYour task is provided as follows:\nText Caption: [virtual dressing room.]" }, { "from": "gpt", "value": "Video 2 is better than Video 1." } ]	[ "losevideos/005028/0.png", "losevideos/005028/1.png", "losevideos/005028/2.png", "losevideos/005028/3.png", "losevideos/005028/4.png", "losevideos/005028/5.png", "losevideos/005028/6.png", "losevideos/005028/7.png", "winvideos/005028/0.png", "winvideos/005028/1.png", "winvideos/005028/2.png", "winvideos/005028/3.png", "winvideos/005028/4.png", "winvideos/005028/5.png", "winvideos/005028/6.png", "winvideos/005028/7.png" ]
American President.	[ { "from": "human", "value": "<image>\n<image>\n<image>\n<image>\n<image>\n<image>\n<image>\n<image>\n <image>\n<image>\n<image>\n<image>\n<image>\n<image>\n<image>\n<image>\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos. You are given a text caption and the frames of two generated videos based on that caption. Your task is to evaluate and compare two videos based on two key criteria:\n1. Alignment with the Caption: Assess how well each video aligns with the provided caption. Consider the accuracy of depicted objects, their relationships, and attributes as described in the caption.\n2. Overall Video Quality: Examine the visual quality of each video, including clarity, detail preservation, color accuracy, and overall aesthetic appeal.\nCompare both videos using the above criteria and select the one that better aligns with the caption while exhibiting superior visual quality.\nProvide a clear conclusion such as \"Video 1 is better than Video 2.\", \"Video 2 is better than Video 1.\" and \"Both videos are equally good.\"\nYour task is provided as follows:\nText Caption: [American President.]" }, { "from": "gpt", "value": "Video 1 is better than Video 2." } ]	[ "winvideos/006742/0.png", "winvideos/006742/1.png", "winvideos/006742/2.png", "winvideos/006742/3.png", "winvideos/006742/4.png", "winvideos/006742/5.png", "winvideos/006742/6.png", "winvideos/006742/7.png", "losevideos/006742/0.png", "losevideos/006742/1.png", "losevideos/006742/2.png", "losevideos/006742/3.png", "losevideos/006742/4.png", "losevideos/006742/5.png", "losevideos/006742/6.png", "losevideos/006742/7.png" ]
two hands holds two gather, a man and a woman's hand, close up, motion 3.	[ { "from": "human", "value": "<image>\n<image>\n<image>\n<image>\n<image>\n<image>\n<image>\n<image>\n <image>\n<image>\n<image>\n<image>\n<image>\n<image>\n<image>\n<image>\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos. You are given a text caption and the frames of two generated videos based on that caption. Your task is to evaluate and compare two videos based on two key criteria:\n1. Alignment with the Caption: Assess how well each video aligns with the provided caption. Consider the accuracy of depicted objects, their relationships, and attributes as described in the caption.\n2. Overall Video Quality: Examine the visual quality of each video, including clarity, detail preservation, color accuracy, and overall aesthetic appeal.\nCompare both videos using the above criteria and select the one that better aligns with the caption while exhibiting superior visual quality.\nProvide a clear conclusion such as \"Video 1 is better than Video 2.\", \"Video 2 is better than Video 1.\" and \"Both videos are equally good.\"\nYour task is provided as follows:\nText Caption: [two hands holds two gather, a man and a woman's hand, close up, motion 3.]" }, { "from": "gpt", "value": "Video 1 is better than Video 2." } ]	[ "winvideos/003137/0.png", "winvideos/003137/1.png", "winvideos/003137/2.png", "winvideos/003137/3.png", "winvideos/003137/4.png", "winvideos/003137/5.png", "winvideos/003137/6.png", "winvideos/003137/7.png", "losevideos/003137/0.png", "losevideos/003137/1.png", "losevideos/003137/2.png", "losevideos/003137/3.png", "losevideos/003137/4.png", "losevideos/003137/5.png", "losevideos/003137/6.png", "losevideos/003137/7.png" ]
Tobuscus wearing a green shirt walking on the rocky shores of ireland at sunrise.	[ { "from": "human", "value": "<image>\n<image>\n<image>\n<image>\n<image>\n<image>\n<image>\n<image>\n <image>\n<image>\n<image>\n<image>\n<image>\n<image>\n<image>\n<image>\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos. You are given a text caption and the frames of two generated videos based on that caption. Your task is to evaluate and compare two videos based on two key criteria:\n1. Alignment with the Caption: Assess how well each video aligns with the provided caption. Consider the accuracy of depicted objects, their relationships, and attributes as described in the caption.\n2. Overall Video Quality: Examine the visual quality of each video, including clarity, detail preservation, color accuracy, and overall aesthetic appeal.\nCompare both videos using the above criteria and select the one that better aligns with the caption while exhibiting superior visual quality.\nProvide a clear conclusion such as \"Video 1 is better than Video 2.\", \"Video 2 is better than Video 1.\" and \"Both videos are equally good.\"\nYour task is provided as follows:\nText Caption: [Tobuscus wearing a green shirt walking on the rocky shores of ireland at sunrise.]" }, { "from": "gpt", "value": "Video 2 is better than Video 1." } ]	[ "losevideos/000813/0.png", "losevideos/000813/1.png", "losevideos/000813/2.png", "losevideos/000813/3.png", "losevideos/000813/4.png", "losevideos/000813/5.png", "losevideos/000813/6.png", "losevideos/000813/7.png", "winvideos/000813/0.png", "winvideos/000813/1.png", "winvideos/000813/2.png", "winvideos/000813/3.png", "winvideos/000813/4.png", "winvideos/000813/5.png", "winvideos/000813/6.png", "winvideos/000813/7.png" ]
Hindu god Cute baby Female on bed natural background, small white flowers, wearing jewelry crown on head, white Skin color, highly detailed, High resolution, High quality, ultra realistic HD,12k.	[ { "from": "human", "value": "<image>\n<image>\n<image>\n<image>\n<image>\n<image>\n<image>\n<image>\n <image>\n<image>\n<image>\n<image>\n<image>\n<image>\n<image>\n<image>\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos. You are given a text caption and the frames of two generated videos based on that caption. Your task is to evaluate and compare two videos based on two key criteria:\n1. Alignment with the Caption: Assess how well each video aligns with the provided caption. Consider the accuracy of depicted objects, their relationships, and attributes as described in the caption.\n2. Overall Video Quality: Examine the visual quality of each video, including clarity, detail preservation, color accuracy, and overall aesthetic appeal.\nCompare both videos using the above criteria and select the one that better aligns with the caption while exhibiting superior visual quality.\nProvide a clear conclusion such as \"Video 1 is better than Video 2.\", \"Video 2 is better than Video 1.\" and \"Both videos are equally good.\"\nYour task is provided as follows:\nText Caption: [Hindu god Cute baby Female on bed natural background, small white flowers, wearing jewelry crown on head, white Skin color, highly detailed, High resolution, High quality, ultra realistic HD,12k.]" }, { "from": "gpt", "value": "Video 2 is better than Video 1." } ]	[ "losevideos/009614/0.png", "losevideos/009614/1.png", "losevideos/009614/2.png", "losevideos/009614/3.png", "losevideos/009614/4.png", "losevideos/009614/5.png", "losevideos/009614/6.png", "losevideos/009614/7.png", "winvideos/009614/0.png", "winvideos/009614/1.png", "winvideos/009614/2.png", "winvideos/009614/3.png", "winvideos/009614/4.png", "winvideos/009614/5.png", "winvideos/009614/6.png", "winvideos/009614/7.png" ]
A climactic scene where Aiden and Elara, standing together, confront a looming catastrophe with their intertwined powers and love, showing their bravery and determination. Message: 1 Attachment.	[ { "from": "human", "value": "<image>\n<image>\n<image>\n<image>\n<image>\n<image>\n<image>\n<image>\n <image>\n<image>\n<image>\n<image>\n<image>\n<image>\n<image>\n<image>\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos. You are given a text caption and the frames of two generated videos based on that caption. Your task is to evaluate and compare two videos based on two key criteria:\n1. Alignment with the Caption: Assess how well each video aligns with the provided caption. Consider the accuracy of depicted objects, their relationships, and attributes as described in the caption.\n2. Overall Video Quality: Examine the visual quality of each video, including clarity, detail preservation, color accuracy, and overall aesthetic appeal.\nCompare both videos using the above criteria and select the one that better aligns with the caption while exhibiting superior visual quality.\nProvide a clear conclusion such as \"Video 1 is better than Video 2.\", \"Video 2 is better than Video 1.\" and \"Both videos are equally good.\"\nYour task is provided as follows:\nText Caption: [A climactic scene where Aiden and Elara, standing together, confront a looming catastrophe with their intertwined powers and love, showing their bravery and determination. Message: 1 Attachment.]" }, { "from": "gpt", "value": "Video 2 is better than Video 1." } ]	[ "losevideos/002262/0.png", "losevideos/002262/1.png", "losevideos/002262/2.png", "losevideos/002262/3.png", "losevideos/002262/4.png", "losevideos/002262/5.png", "losevideos/002262/6.png", "losevideos/002262/7.png", "winvideos/002262/0.png", "winvideos/002262/1.png", "winvideos/002262/2.png", "winvideos/002262/3.png", "winvideos/002262/4.png", "winvideos/002262/5.png", "winvideos/002262/6.png", "winvideos/002262/7.png" ]
a craftsman.	[ { "from": "human", "value": "<image>\n<image>\n<image>\n<image>\n<image>\n<image>\n<image>\n<image>\n <image>\n<image>\n<image>\n<image>\n<image>\n<image>\n<image>\n<image>\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos. You are given a text caption and the frames of two generated videos based on that caption. Your task is to evaluate and compare two videos based on two key criteria:\n1. Alignment with the Caption: Assess how well each video aligns with the provided caption. Consider the accuracy of depicted objects, their relationships, and attributes as described in the caption.\n2. Overall Video Quality: Examine the visual quality of each video, including clarity, detail preservation, color accuracy, and overall aesthetic appeal.\nCompare both videos using the above criteria and select the one that better aligns with the caption while exhibiting superior visual quality.\nProvide a clear conclusion such as \"Video 1 is better than Video 2.\", \"Video 2 is better than Video 1.\" and \"Both videos are equally good.\"\nYour task is provided as follows:\nText Caption: [a craftsman.]" }, { "from": "gpt", "value": "Video 1 is better than Video 2." } ]	[ "winvideos/006436/0.png", "winvideos/006436/1.png", "winvideos/006436/2.png", "winvideos/006436/3.png", "winvideos/006436/4.png", "winvideos/006436/5.png", "winvideos/006436/6.png", "winvideos/006436/7.png", "losevideos/006436/0.png", "losevideos/006436/1.png", "losevideos/006436/2.png", "losevideos/006436/3.png", "losevideos/006436/4.png", "losevideos/006436/5.png", "losevideos/006436/6.png", "losevideos/006436/7.png" ]
flying over a coconut beach -ar 16:9 -mo 4.	[ { "from": "human", "value": "<image>\n<image>\n<image>\n<image>\n<image>\n<image>\n<image>\n<image>\n <image>\n<image>\n<image>\n<image>\n<image>\n<image>\n<image>\n<image>\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos. You are given a text caption and the frames of two generated videos based on that caption. Your task is to evaluate and compare two videos based on two key criteria:\n1. Alignment with the Caption: Assess how well each video aligns with the provided caption. Consider the accuracy of depicted objects, their relationships, and attributes as described in the caption.\n2. Overall Video Quality: Examine the visual quality of each video, including clarity, detail preservation, color accuracy, and overall aesthetic appeal.\nCompare both videos using the above criteria and select the one that better aligns with the caption while exhibiting superior visual quality.\nProvide a clear conclusion such as \"Video 1 is better than Video 2.\", \"Video 2 is better than Video 1.\" and \"Both videos are equally good.\"\nYour task is provided as follows:\nText Caption: [flying over a coconut beach -ar 16:9 -mo 4.]" }, { "from": "gpt", "value": "Video 2 is better than Video 1." } ]	[ "losevideos/008570/0.png", "losevideos/008570/1.png", "losevideos/008570/2.png", "losevideos/008570/3.png", "losevideos/008570/4.png", "losevideos/008570/5.png", "losevideos/008570/6.png", "losevideos/008570/7.png", "winvideos/008570/0.png", "winvideos/008570/1.png", "winvideos/008570/2.png", "winvideos/008570/3.png", "winvideos/008570/4.png", "winvideos/008570/5.png", "winvideos/008570/6.png", "winvideos/008570/7.png" ]
Show scenes where Aelia encounters various people and creatures, learning wisdom and growing with each encounter.	[ { "from": "human", "value": "<image>\n<image>\n<image>\n<image>\n<image>\n<image>\n<image>\n<image>\n <image>\n<image>\n<image>\n<image>\n<image>\n<image>\n<image>\n<image>\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos. You are given a text caption and the frames of two generated videos based on that caption. Your task is to evaluate and compare two videos based on two key criteria:\n1. Alignment with the Caption: Assess how well each video aligns with the provided caption. Consider the accuracy of depicted objects, their relationships, and attributes as described in the caption.\n2. Overall Video Quality: Examine the visual quality of each video, including clarity, detail preservation, color accuracy, and overall aesthetic appeal.\nCompare both videos using the above criteria and select the one that better aligns with the caption while exhibiting superior visual quality.\nProvide a clear conclusion such as \"Video 1 is better than Video 2.\", \"Video 2 is better than Video 1.\" and \"Both videos are equally good.\"\nYour task is provided as follows:\nText Caption: [Show scenes where Aelia encounters various people and creatures, learning wisdom and growing with each encounter.]" }, { "from": "gpt", "value": "Video 2 is better than Video 1." } ]	[ "losevideos/004629/0.png", "losevideos/004629/1.png", "losevideos/004629/2.png", "losevideos/004629/3.png", "losevideos/004629/4.png", "losevideos/004629/5.png", "losevideos/004629/6.png", "losevideos/004629/7.png", "winvideos/004629/0.png", "winvideos/004629/1.png", "winvideos/004629/2.png", "winvideos/004629/3.png", "winvideos/004629/4.png", "winvideos/004629/5.png", "winvideos/004629/6.png", "winvideos/004629/7.png" ]
Iron Man flies in the sky.	[ { "from": "human", "value": "<image>\n<image>\n<image>\n<image>\n<image>\n<image>\n<image>\n<image>\n <image>\n<image>\n<image>\n<image>\n<image>\n<image>\n<image>\n<image>\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos. You are given a text caption and the frames of two generated videos based on that caption. Your task is to evaluate and compare two videos based on two key criteria:\n1. Alignment with the Caption: Assess how well each video aligns with the provided caption. Consider the accuracy of depicted objects, their relationships, and attributes as described in the caption.\n2. Overall Video Quality: Examine the visual quality of each video, including clarity, detail preservation, color accuracy, and overall aesthetic appeal.\nCompare both videos using the above criteria and select the one that better aligns with the caption while exhibiting superior visual quality.\nProvide a clear conclusion such as \"Video 1 is better than Video 2.\", \"Video 2 is better than Video 1.\" and \"Both videos are equally good.\"\nYour task is provided as follows:\nText Caption: [Iron Man flies in the sky.]" }, { "from": "gpt", "value": "Video 1 is better than Video 2." } ]	[ "winvideos/009704/0.png", "winvideos/009704/1.png", "winvideos/009704/2.png", "winvideos/009704/3.png", "winvideos/009704/4.png", "winvideos/009704/5.png", "winvideos/009704/6.png", "winvideos/009704/7.png", "losevideos/009704/0.png", "losevideos/009704/1.png", "losevideos/009704/2.png", "losevideos/009704/3.png", "losevideos/009704/4.png", "losevideos/009704/5.png", "losevideos/009704/6.png", "losevideos/009704/7.png" ]
fast zoom in and out, cyberpunk girl dancing in ecstasy in disco club, practicing enchanting magic, laser show, sensually moving her body, sparkling lights, ray of lights, depth of fields, hyper-realistic render, cinematic view style, 4K -ar 16:9 -motion 2 -gs 15.	[ { "from": "human", "value": "<image>\n<image>\n<image>\n<image>\n<image>\n<image>\n<image>\n<image>\n <image>\n<image>\n<image>\n<image>\n<image>\n<image>\n<image>\n<image>\nSuppose you are an expert in judging and evaluating the quality of AI-generated videos. You are given a text caption and the frames of two generated videos based on that caption. Your task is to evaluate and compare two videos based on two key criteria:\n1. Alignment with the Caption: Assess how well each video aligns with the provided caption. Consider the accuracy of depicted objects, their relationships, and attributes as described in the caption.\n2. Overall Video Quality: Examine the visual quality of each video, including clarity, detail preservation, color accuracy, and overall aesthetic appeal.\nCompare both videos using the above criteria and select the one that better aligns with the caption while exhibiting superior visual quality.\nProvide a clear conclusion such as \"Video 1 is better than Video 2.\", \"Video 2 is better than Video 1.\" and \"Both videos are equally good.\"\nYour task is provided as follows:\nText Caption: [fast zoom in and out, cyberpunk girl dancing in ecstasy in disco club, practicing enchanting magic, laser show, sensually moving her body, sparkling lights, ray of lights, depth of fields, hyper-realistic render, cinematic view style, 4K -ar 16:9 -motion 2 -gs 15.]" }, { "from": "gpt", "value": "Video 2 is better than Video 1." } ]	[ "losevideos/004287/0.png", "losevideos/004287/1.png", "losevideos/004287/2.png", "losevideos/004287/3.png", "losevideos/004287/4.png", "losevideos/004287/5.png", "losevideos/004287/6.png", "losevideos/004287/7.png", "winvideos/004287/0.png", "winvideos/004287/1.png", "winvideos/004287/2.png", "winvideos/004287/3.png", "winvideos/004287/4.png", "winvideos/004287/5.png", "winvideos/004287/6.png", "winvideos/004287/7.png" ]