MenghaoGuo / AutoDeeplab

Pytorch Implementation the paper Auto-DeepLab Hierarchical Neural Architecture Search for Semantic Image Segmentation
https://arxiv.org/abs/1901.02985
410 stars 97 forks source link

Given groups=1, weight of size 48 160 1 1, expected input[2, 1280, 25, 25] to have 160 channels, but got 1280 channels instead #62

Open mx1mx2 opened 4 years ago