search

wei-tim / YOWO

You Only Watch Once: A Unified CNN Architecture for Real-Time Spatiotemporal Action Localization
Other
846 stars 158 forks source link

how to input a video and outputs a video in the format of ava #76

Closed xuhui1994 closed 3 years ago

okankop commented 3 years ago

Please check the file test_video_ava.py

xuhui1994 commented 3 years ago

Thank your reply, I had been checked test_video_ava.py before. This code just outputs images whose size is 224x224. I wanna to get the predict video  whose resolution is same as test video (for example,1920x1080),rather than 224x224 images 

------------------ 原始邮件 ------------------ 发件人: "wei-tim/YOWO" @.>; 发送时间: 2021年4月20日(星期二) 下午5:22 @.>; @.**@.>; 主题: Re: [wei-tim/YOWO] how to input a video and outputs a video in the format of ava (#76)

Closed #76.

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub, or unsubscribe.

okankop commented 3 years ago

You can resize your video to 224x224 get your predictions and resize these predictions back to 1920x1080.

If you want to use resolution 1920x1080 directly at inference, you need to train the models again with this resolution.