YeTaoY
YeTaoY
> > Finally, I modified the training_hooks configuration of the train.yaml file, and the changes are as follows: > > ``` > > training_hooks: > > lr_config: > > policy:...
I have the same problem on training Kinetics data. loss is about 6, and can not converge
Hi, guys , what @vivek87799 has done fixed my problem.
@gavenswang 你测试的时候用的.caffemodel还是.npz
@gavenswang caffemodel并没有改变模型的大小，虽然已经实现的量化，但是caffe的保存用的还是fp32的，真正被替换过存储方式的只有npz. 我好奇的是npz应该怎么用，可否推荐个使用方法（把它decode成caffemodel在用caffe测的方式不是我需要的）