cannot use the pretrain model

Pongpisit-Thanasutives / Variations-of-SFANet-for-Crowd-Counting

The official implementation of "Encoder-Decoder Based Convolutional Neural Networks with Multi-Scale-Aware Modules for Crowd Counting"

https://ieeexplore.ieee.org/document/9413286

GNU General Public License v3.0

110 stars 32 forks source link

cannot use the pretrain model #9

Closed ckqsars closed 4 years ago

ckqsars commented 4 years ago

I want to use the pretain model to test the Visualization code. But have erros in loading state_dict for Model: the error is miss key of vgg's weight and bias. So How can I use the pretrain model?

Pongpisit-Thanasutives commented 4 years ago

I think you need the pretrained weights of M_SegNet to test the visualization code. Here it is https://drive.google.com/file/d/1fGuH4o0hKbgdP1kaj9rbjX2HUL1IH0oo/view?usp=sharing. Note that the weights for SHA and SHB are for M_SegNet (not M_SegNet).