search

zhang-tao-whu / DVIS

DVIS: Decoupled Video Instance Segmentation Framework
MIT License
124 stars 7 forks source link

VSS bug #5

Closed wengkai0 closed 1 year ago

wengkai0 commented 1 year ago

Hi, authors,

Thanks for your great work! I met a few problems when running VSS code, could you pls give any suggestions?

  1. How do I train DVIS on VSS? (following which config file?)
  2. and What is the difference between 480p and 720p dataset?

I follow the code python train_net_video.py --num-gpus 4 --config-file configs/VSPW/MinVIS_R50_480p.yaml

The traceback error is : sem_seg_gt[sem_seg_gt == 0] = 255 ValueError: assignment destination is read-only

Thanks!

zhang-tao-whu commented 1 year ago

Sorry, I will check and fix this bug as soon as possible.

wengkai0 commented 1 year ago

Thanks!

I also found that the inference code for VPS and VSS does not work. Since the current code does not specify that it is one of the tasks in VSS or VPS. Could you please have a check?

zhang-tao-whu commented 1 year ago

Could you provide your config file? cfg.MODEL.MASK_FORMER.TEST.TASK specifies the specified inference task ('vis', 'vss'or 'vps') for the network. Please refer to line 22 of config.

wengkai0 commented 1 year ago

Hi Zhangtao,

Thanks for your information, I will have a try and give feedback to you.

And could I put more than one samples in one GPU to train DVIS?

Thanks

zhang-tao-whu commented 1 year ago

I'm sorry, currently we only support one GPU per batch of video data. However, there's a possibility that in the future, we may support multiple batches of video for a single GPU.

wengkai0 commented 1 year ago

Thanks for your prompt reply, It helps a lot!