carrierlxk / AGNN

Zero-shot Video Object Segmentation via Attentive Graph Neural Networks (ICCV2019 Oral)
207 stars 25 forks source link

请问为什么要对batch做循环 #15

Open fupiao1998 opened 2 years ago

fupiao1998 commented 2 years ago

您好,非常感谢您开源代码。在代码的模型部分,我看到主模型中的forward对每一个batch做了循环,每一次循环独立处理,在我理解中,batch本身就是独立的,那么是否可以不做循环而直接批次处理呢?

carrierlxk commented 2 years ago

你好,

我觉得你可以试试。应该影响不是很大

fupiao1998 @.***> 于2022年5月3日周二 14:26写道:

您好,非常感谢您开源代码。在代码的模型部分,我看到主模型中的forward对每一个batch做了循环,每一次循环独立处理,在我理解中,batch本身就是独立的,那么是否可以不做循环而直接批次处理呢?

— Reply to this email directly, view it on GitHub https://github.com/carrierlxk/AGNN/issues/15, or unsubscribe https://github.com/notifications/unsubscribe-auth/ABNY6ZU2TK3CXO2VW4RMX7LVIDBIFANCNFSM5U57TRVA . You are receiving this because you are subscribed to this thread.Message ID: @.***>

-- Xiankai Lu (卢宪凯), Research Professor Research Center of Artificial Intelligence School of Software Shandong University, 250101, Jinan, China

fupiao1998 commented 2 years ago

感谢回复!