huyi

> I set batch_size to 48, and the program works, but set batch_size > 48, the above error will appear. Do you solve it?

> > Thanks for the fruitful discussion. > > I got an accuracy of 94% on UCSDped2 using the pretrained model ckpt, the only difference I had with @Wolfybox 's...

> My re-implemented results in unsupervised anomaly detection (1000 epochs for all the three datasets): **Ped2: 95.0% Avenue: 82.2% ShanghaiTech: 72.9%** The result on Avenue is much lower than that...

> > 话说你们是怎么训练的呀？是直接就作者的代码，训练1000epoch么？我看论文写的是epoch=60，我跑完了之后，ped2只有95.1。低很多。大道也是。 > > 我跑了一下，按照提示参数也都修改了，结果测试提示outputs, feas, _, _, _, fea_loss = model.forward(imgs[:,:3*4], update_weights, False) ValueError: not enough values to unpack (expected 6, got 2)，有没有什么办法解决 请问你解决了么