hustvl/MIMDet

Issues

• can‘t ’get 51.8 AP on coco use the source code

  #29 opened 6 months ago by UcanSee
  1

• Distributed package

  #28 opened 10 months ago by RileyMccon
  0

• Thanks for sharing the visualization code

  #27 opened 10 months ago by QY1994-0919
  0

• FloatingPointError: Predicted boxes or scores contain Inf/NaN. Training has diverged.

  #26 opened a year ago by mike-huangdj
  1

• Inf/NaN error happens during the training

  #7 opened 3 years ago by junchen14
  6

• question about the randomly initialized patch embedding layer

  #25 opened 2 years ago by ydhongHIT
  0

• How to visualize results？

  #22 opened 2 years ago by LowPower-Center
  3

• Possible to remove Detectron2 like YOLOS?

  #24 opened 2 years ago by jordanparker6
  1

• RuntimeError: CUDA out of memory during training, works fine during inference.

  #23 opened 3 years ago by sfarkya
  7

• Error in running inference.

  #20 opened 3 years ago by sfarkya
  7

• No predictions from model

  #21 opened 3 years ago by normster
  3

• how to re-produce the result with only 8-GPUs

  #3 opened 3 years ago by youngwanLEE
  8

• How should I set the dataset's path??

  #18 opened 3 years ago by ucasligang
  2

• Could you provide the code to extract the object embedding?

  #17 opened 3 years ago by Pter61
  0

• I want to do the downstream detection task based on your upstream pretraining model. Do you have a more detailed document

  #8 opened 3 years ago by raychiu0202
  2

• Some implementation detail questions

  #14 opened 3 years ago by Yingdong-Hu
  2

• About sincos_pos_embed

  #15 opened 3 years ago by qiy20
  5

• some problems in runtime

  #16 opened 3 years ago by wudizuixiaosa
  0

• How to change the config to train Benchmarking-ViT-B with batch size 16 ?

  #13 opened 3 years ago by Yingdong-Hu
  2

• Hi, what is the writing method to the right of the equal sign in statement ' img_size="${..encoder.img_size}" '

  #12 opened 3 years ago by ross-Hr
  3

• Loading the pre-trained weight from MAE

  #11 opened 3 years ago by youngwanLEE
  3

• MIM training while training object detection

  #10 opened 3 years ago by bamps53
  2

• multi-machine training in clustre

  #4 opened 3 years ago by xushilin1
  1

• AP performance is not consistent with training log?

  #9 opened 3 years ago by zhanghengdev
  2

• problem about config of benchmark

  #5 opened 3 years ago by UcanSee
  2

• Could you provide a model of 768 dimension output image features?

  #6 opened 3 years ago by Pter61
  2

• add web demo/model to Huggingface

  #2 opened 3 years ago by AK391
  2

• does it work without finetuning directly using pretrained model for detection?

  #1 opened 3 years ago by lucasjinreal
  4