This is the extent dataset proposed in the paper "OpenVid-1M: A Large-Scale High-Quality Dataset for Text-to-video Generation". OpenVid-1M is a high-quality text-to-video dataset designed for research institutions to enhance video quality, featuring high aesthetics, clarity, and resolution. It can be used for direct training or as a quality tuning complement to other video datasets.
New Feature: Video-ZIP mapping files now available for efficient video lookup (see Dataset Structure and Extended Usage).
DATA_PATH
└── video_mappings (NEW!)
├── OpenVid_part0.csv
├── OpenVid_part1.csv
├── ...
└── OpenVid_part185.csv
├── README.md
└── ...
Video-ZIP Mapping
We provide 186 mapping files (OpenVid_part*.csv) containing:
zip_file,video,video_path
OpenVid_part0.zip,celebv_--Jiv5iYqT8_0.mp4,OpenVid_part1/celebv_--Jiv5iYqT8_0.mp4
OpenVid_part0.zip,celebv_--Jiv5iYqT8_2.mp4,OpenVid_part1/celebv_--Jiv5iYqT8_2.mp4
OpenVid_part0.zip,celebv_--QCZKgJt6o_0.mp4,OpenVid_part1/celebv_--QCZKgJt6o_0.mp4
OpenVid_part0.zip,celebv_--oCWVOBuvA_0.mp4,OpenVid_part1/celebv_--oCWVOBuvA_0.mp4
...
zip_file: Source ZIP archive namevideo: Video filenamevideo_path: Relative path within the ZIP fileMapping Features
import pandas as pd
# Load all mapping files
mappings = [pd.read_csv(f"video_mappings/OpenVid_part{i}.csv")
for i in range(186)]
full_mapping = pd.concat(mappings)
# Find ZIP location for specific video
video_name = "celebv_--Jiv5iYqT8_0.mp4"
result = full_mapping[full_mapping.video == video_name]
print(f"Video {video_name} is in {result.zip_file.values[0]} at path {result.video_path.values[0]}")
# Extract specific video from ZIP without full decompression
unzip OpenVid_part0.zip "OpenVid_part1/celebv_--Jiv5iYqT8_0.mp4" -d target_folder
# Unzip files (-j flattens directory structure)
This mapping dataset was developed to enhance OpenVid-1M accessibility. Special thanks to the original dataset contributors.
@article{nan2024openvid,
title={OpenVid-1M: A Large-Scale High-Quality Dataset for Text-to-video Generation},
author={Nan, Kepan and Xie, Rui and Zhou, Penghao and Fan, Tiehan and Yang, Zhenheng and Chen, Zhijie and Li, Xiang and Yang, Jian and Tai, Ying},
journal={arXiv preprint arXiv:2407.02371},
year={2024}
}
This is the extent dataset proposed in the paper "OpenVid-1M: A Large-Scale High-Quality Dataset for Text-to-video Generation". OpenVid-1M is a high-quality text-to-video dataset designed for research institutions to enhance video quality, featuring high aesthetics, clarity, and resolution. It can be used for direct training or as a quality tuning complement to other video datasets.
New Feature: Video-ZIP mapping files now available for efficient video lookup (see Dataset Structure and Extended Usage).
DATA_PATH
└── video_mappings (NEW!)
├── OpenVid_part0.csv
├── OpenVid_part1.csv
├── ...
└── OpenVid_part185.csv
├── README.md
└── ...
Video-ZIP Mapping
We provide 186 mapping files (OpenVid_part*.csv) containing:
zip_file,video,video_path
OpenVid_part0.zip,celebv_--Jiv5iYqT8_0.mp4,OpenVid_part1/celebv_--Jiv5iYqT8_0.mp4
OpenVid_part0.zip,celebv_--Jiv5iYqT8_2.mp4,OpenVid_part1/celebv_--Jiv5iYqT8_2.mp4
OpenVid_part0.zip,celebv_--QCZKgJt6o_0.mp4,OpenVid_part1/celebv_--QCZKgJt6o_0.mp4
OpenVid_part0.zip,celebv_--oCWVOBuvA_0.mp4,OpenVid_part1/celebv_--oCWVOBuvA_0.mp4
...
zip_file: Source ZIP archive namevideo: Video filenamevideo_path: Relative path within the ZIP fileMapping Features
import pandas as pd
# Load all mapping files
mappings = [pd.read_csv(f"video_mappings/OpenVid_part{i}.csv")
for i in range(186)]
full_mapping = pd.concat(mappings)
# Find ZIP location for specific video
video_name = "celebv_--Jiv5iYqT8_0.mp4"
result = full_mapping[full_mapping.video == video_name]
print(f"Video {video_name} is in {result.zip_file.values[0]} at path {result.video_path.values[0]}")
# Extract specific video from ZIP without full decompression
unzip OpenVid_part0.zip "OpenVid_part1/celebv_--Jiv5iYqT8_0.mp4" -d target_folder
# Unzip files (-j flattens directory structure)
This mapping dataset was developed to enhance OpenVid-1M accessibility. Special thanks to the original dataset contributors.
@article{nan2024openvid,
title={OpenVid-1M: A Large-Scale High-Quality Dataset for Text-to-video Generation},
author={Nan, Kepan and Xie, Rui and Zhou, Penghao and Fan, Tiehan and Yang, Zhenheng and Chen, Zhijie and Li, Xiang and Yang, Jian and Tai, Ying},
journal={arXiv preprint arXiv:2407.02371},
year={2024}
}