jyli

如何处理的student多次forward 的loss

2

您好，请问代码中student model对于有标签书先进行了forwardtrain得到了losses，对于无标签数据又单独进行了forwardtrain，这样两次forward不会导致student网络中同一个参数点在计算图中被标记了两次，会导致loss.backward报错吗？

File "/home/xxx/workspace/ONCE_Benchmark/pcdet/ops/dcn/deform_conv.py", line 87, in backward deform_conv_cuda.deform_conv_backward_parameters_cuda( RuntimeError: view size is not compatible with input tensor's size and stride (at least one dimension spans across two contiguous subspaces). Use .reshape(...)...

咨询一些teacher update的具体细节？

1

您好， 感谢你提出的优秀的方法，我有一些小疑问： 论文中实验细节部分说半监督阶段学习率是恒定的0.001，并且教师模型每25个epoch更新一次：Then the student is trained for 25 to 75 epochs depending on the amount of unlabeled data with learning rate 0.001, and the teacher is updated every 25...