leviswind

icon company Shanghai Jiao Tong University icon location Shanghai, China icon location Phd in SJTU

Results 2 issues of leviswind

关于dataSupport_qa.py

1 comment

假设training的样本数为1000，image就是1000*3*224*224，我想问一下179行这里，如果bucket数目不为1，那么这个idx就不是原来的idx，因为len(x)小于1000，是不是这里有问题。当然如果bucket数目为1，这里就没有问题。

train from scratch on ucf101 dataset

10 comment

We try to train i3d model on ucf101 from scratch, but it converges much slower with a final validation accuray around 60%. Can you offer some suggestions on train i3d...