XinWangg
XinWangg
@khawar512 Yes, It takes a lot of time to train, and when I train the model, I found that the loss may not drop as quickly as CNN. So I...
> Please refer to the training logs for ViT-P10S8. I have uploaded a 'log' folder. Thanks!
@zhongyy Can you share your training log?
@khawar512 Have you reproduce the result?
你训练几个epoch呢？
我训练了120个epoch，batchsize设为64，loss到6.多就不降了。 | | 王鑫 | | 邮箱：[email protected] | 签名由 网易邮箱大师 定制 在2019年11月06日 19:24，winfred 写道： 200个epoch,我的batch_size设置的32,5的话太不稳定了 ------------------ 原始邮件 ------------------ 发件人: "XinWang-bip"
Loss3.62的效果怎样，我用loss6的模型测试了voc2007效果很差 | | 王鑫 | | 邮箱：[email protected] | 签名由 网易邮箱大师 定制 在2019年11月06日 19:32，winfred 写道： 我到3.62就开始跳动了，下不去了。给更小的学习率也没用 — You are receiving this because you commented. Reply to this email directly, view it...
[dpcrn_hop512.zip](https://github.com/Tencent/ncnn/files/13574600/dpcrn_hop512.zip)
希望能得到您的解答！
@yangyyt 请问有找到比较好的解决方法没~