English
全部
图片
视频
地图
资讯
购物
更多
航班
旅游
酒店
搜索
笔记本
Top stories
Sports
U.S.
Local
World
Science
Technology
Entertainment
Business
More
Politics
时间不限
过去 1 小时
过去 24 小时
过去 7 天
过去 30 天
按相关度排序
按时间排序
腾讯网
9 天
日本团队分享基于预训练VQGAN的高效全景图像合成方法2S-ODIS
针对相关问题,日本 上智大学 团队提出了一种新的全向图像合成方法2S-ODIS ,它不仅可以生成高质量的全向图像,同时大大缩短了训练时间。这是通过使用在大规模NFoV图像数据库(如ImageNet)预训练的Vector Quantized GAN模型而不进行微调来实现。
当前正在显示可能无法访问的结果。
隐藏无法访问的结果
今日热点
Guilty of murder, hate crime
Tensions flare at meeting
LSU athlete dies in crash
US OKs arms sale to Israel
$81T credited by mistake
Red flag warnings issued
Task force to visit campuses
Ammunition plant explosion
I-40 to partially reopen
Tapped for Navy’s No. 2 post
Delays $28B Ohio project
Drug lord pleads not guilty
To shut down in May
‘Police Story' creator dies
FBI returns Trump materials
Judge blocks Trump order
Carbon monoxide ruled out
Drake settles legal dispute
FedEx plane catches fire
Minneapolis man charged
Embiid ruled out for season
AZ plane crash report
FAA OKs Starship Flight 8
NASA set to launch SPHEREx
Stabbing suspect arrested
To cut about 7,000 workers
TX measles outbreak grows
DNC sues over Trump order
Drops enforcement suits
反馈