search

bowenc0221 / panoptic-deeplab

This is Pytorch re-implementation of our CVPR 2020 paper "Panoptic-DeepLab: A Simple, Strong, and Fast Baseline for Bottom-Up Panoptic Segmentation" (https://arxiv.org/abs/1911.10194)
Apache License 2.0
590 stars 117 forks source link

pool_kernal_size error! #64

Closed moodlife closed 3 years ago

moodlife commented 3 years ago

When I want to change the crop size of coco_dataset according to your config file, I found that there is an unexpected error. pool__kernal_size must be divisible by the shape of inputs Input size:[40:80] pool_kernal_size[80,80]

moodlife commented 3 years ago

And I try to trace the input image but find it wasnot cropped in the correct size. However, when I set the crop size to 640*640, it works.

moodlife commented 3 years ago

I wonder if you have achieved the crop with pad? I think this error is about crop You have mentioned that D2-version supported the pad, but I can't find its part. I wonder if you can give some advice Sincerely!