NAN value when using the pre-trained model

micmic123 / QmapCompression

Official implementation of "Variable-Rate Deep Image Compression through Spatially-Adaptive Feature Transform", ICCV 2021

81 stars 16 forks source link

Open APiaoG opened 1 year ago

APiaoG commented 1 year ago

Hello! Great works! What is the reason why I encountered a nan value when using the pre trained model? Looking forward to your reply!

Kei-Iino commented 1 year ago

same situation now

jinhaohust commented 2 months ago

same situation now, I don't know if you have solved this problem.