Sense-X/X-Temporal

Issues

• Can this project continue to be updated

  #23 opened 3 years ago by NEUdeep
  0

• without close tensorboard, there will be a EOFError at the end of training

  #22 opened 3 years ago by wwnbbd
  1

• How can you get the meta files: .tex > Thanks for your feedback. You can set trainer.no_partial_bn = True if batch size >= 6 in each gpu and retry it, this will not affect the accuracy. That module exists some bug with distributed training, we will fix it quickly.

  #17 opened 3 years ago by SceneRec
  1

• `AverageMeter` is not correct

  #19 opened 3 years ago by CarolineCheng233
  1

• When training，the log file stop at 'save_dir: checkpoint/' and with no update again

  #20 opened 3 years ago by SceneRec
  1

• Found Two fresh BUGs and a Solution

  #21 opened 3 years ago by SceneRec
  1

• issues with pytorch version higher than 1.4.0

  #18 opened 4 years ago by lininglouis
  1

• Prediction demo using different pretrained models

  #16 opened 4 years ago by kshitijd20
  2

• About the model result

  #15 opened 4 years ago by wwnbbd
  4

• Pretrained model config file

  #14 opened 4 years ago by wwnbbd
  5

• Abort the multi-label

  #13 opened 4 years ago by simobupt
  4

• RuntimeError: DataLoader worker (pid(s) 1995171, 1996371) exited unexpectedly

  #12 opened 4 years ago by zeng-lingyun
  2

• Loss is zero when using single GPU

  #11 opened 4 years ago by JaywongWang
  0

• Feature extraction

  #9 opened 4 years ago by Finspire13
  1

• forkserver

  #7 opened 4 years ago by haooooooqi
  1

• how to get the TIN pretrain model in Kinetics-600?

  #8 opened 4 years ago by MillX2021
  1

• evaluate on test dataset when training?

  #4 opened 4 years ago by HC-2016
  7

• In linear_sampler(data, bias), bias 'shape should N*T*G?

  #6 opened 4 years ago by cvnovice95
  1

• meta_file when training?

  #5 opened 4 years ago by simobupt
  0

• Could you release trained models?

  #3 opened 4 years ago by HC-2016
  2

• Can I run without GPU?

  #2 opened 4 years ago by Li-Shu14
  1

• ./train.sh for TSM stop at the first log infor. : Freezing BatchNorm2D except...

  #1 opened 4 years ago by Amazingren
  6