search

zhang-tao-whu / DVIS

DVIS: Decoupled Video Instance Segmentation Framework
MIT License
127 stars 6 forks source link

vspw inference results image pixels is all zero #31

Closed steven-spec closed 5 months ago

steven-spec commented 5 months ago

the command use for inference: python train_net_video.py \ --num-gpus 8 \ --config-file configs/SDG/DVIS_Online_R50_720p.yaml \ --eval-only MODEL.WEIGHTS output_DVIS_Online_R50_720p_VSPW/model_final.pth

train command: python train_net_video.py --num-gpus 8 --config-file configs/VSPW/DVIS_Online_R50_720p.yaml \

zhang-tao-whu commented 5 months ago

What is SDG? Can you share the config file (configs/SDG/DVIS_Online_R50_720p.yaml ) and the training and testing logs to help me understand your problems?

steven-spec commented 5 months ago

What is SDG? Can you share the config file (configs/SDG/DVIS_Online_R50_720p.yaml ) and the training and testing logs to help me understand your problems?

Sorry, I copied the wrong path， the correct config file path is :configs/VSPW/DVIS_Online_R50_720p.yaml. I have retrained the model and it can now inference normally, thank you for your reply!

I would like to know if there has been any progress in your response to issue:#14