search

shunsukesaito / PIFu

This repository contains the code for the paper "PIFu: Pixel-Aligned Implicit Function for High-Resolution Clothed Human Digitization"
https://shunsukesaito.github.io/PIFu/
Other
1.76k stars 341 forks source link

trian issue #58

Closed jiandandian2 closed 4 years ago

jiandandian2 commented 4 years ago

hello, i used the training as your readme saying, and i test using the trained model : netG_epoch_99 and net_C_epoch_99, but it cannot work image

shunsukesaito commented 4 years ago

Looks like your image filter of ResBlkPIFuNet during training and inference are not compatible. Please make sure you use the same model for both.

Henry831 commented 4 years ago

I have a same problem,how can i get the model like your trained model?

jiandandian2 commented 4 years ago

I have a same problem,how can i get the model like your trained model?

I do as readme saying, data generation(step1 and step2 generate the standard dataset) and training(step1 output shape_weights, step2 output color weights)

Henry831 commented 4 years ago

I have a same problem,how can i get the model like your trained model?

I do as readme saying, data generation(step1 and step2 generate the standard dataset) and training(step1 output shape_weights, step2 output color weights)

Yes,I have a same process like yours

shunsukesaito commented 4 years ago

Can you check if #61 solves your issue as well?

jiandandian2 commented 4 years ago

Can you check if #61 solves your issue as well? it didn't , I train and test under the same environment...

jiandandian2 commented 4 years ago

61 sovled the problem