liuzhifeng

我在训练的时候也越大维度不匹配的问题。 `dice += (pred[:,i] * target[:,i]).sum(dim=1).sum(dim=1).sum(dim=1) / ((pred[:,i] * target[:,i]).sum(dim=1).sum(dim=1).sum(dim=1)+ RuntimeError: The size of tensor a (512) must match the size of tensor b (256) at non-singleton dimension 3`

Thank you very much for your reply ! I just want to recurrence your results which appear in the paper.

> Hello. To run a the pipeline with a 3D model, one would use the [3d inference config](https://github.com/mehta-lab/microDL/blob/master/config_files/config_inference_3d.yml), [2.5d preprocessing](https://github.com/mehta-lab/microDL/blob/master/config_files/Preprocessing-config_2.5DUnet_regression_phase2membrane.yml), and slightly modify one of the [training configs](https://github.com/mehta-lab/microDL/blob/master/config_files/Training-config_2.5DUnet_regression_phase2membrane.yml) to fit...

I want to reproduce the resluts of **Figure 3** in your paper, which include 3D Predicted F-actin.

> And I want to know that why "Depth must be uneven" in **def adjust_slice_margins(slice_ids,depth)** of [aux_utils.py](https://github.com/mehta-lab/microDL/blob/master/micro_dl/utils/aux_utils.py)