taoyang1122/adapt-image-models

Issues

• Test results

  #48 opened 7 months ago by Liu-arch
  0

• model下载问题

  #47 opened 10 months ago by Douyi2333
  0

• I would like to know what testing protocol the 88.9% on Diving48 is based on.

  #46 opened a year ago by Changwei-Ouyang
  0

• Have the authors tried fine-tuning the parameters of the LayerNorm layer with it turned on? If so, what were the results?

  #42 opened a year ago by Changwei-Ouyang
  1

• An error with training on multiple GPUs

  #45 opened a year ago by wlsrick
  1

• visualizing code

  #41 opened a year ago by Changwei-Ouyang
  0

• A Error in the Running the Train Code.

  #44 opened a year ago by YeungChiu
  0

• The issue of being unable to continue training with already trained weights.

  #43 opened a year ago by Xfold-masker
  0

• How to finetune?

  #40 opened a year ago by smandava98
  1

• The training set and val set in the Diving48 dataset

  #39 opened a year ago by Changwei-Ouyang
  1

• The dimension of vit_clip and vit_imagenet

  #38 opened a year ago by wlsrick
  2

• About the performance gap with the released checkpoints

  #35 opened a year ago by yangbang18
  6

• About learning rate setting

  #37 opened a year ago by hsi-che-lin
  1

• pretrained model

  #19 opened 2 years ago by dmmSJTU
  7

• Can't reproduce Something-Somethingv2 training and results

  #25 opened 2 years ago by or7r
  7

• Weights Initialization

  #36 opened a year ago by liyi-ff
  1

• About hyperparameters setting for reproducing the results

  #32 opened a year ago by hsi-che-lin
  3

• Train the ViT in 4 2080Ti

  #34 opened a year ago by wlsrick
  1

• Encoding error while loading val data on k400

  #33 opened a year ago by TJQdoIt9527
  1

• Inquiring about the number of views during test time.

  #31 opened a year ago by Backdrop9019
  1

• About GPU memory usage

  #30 opened a year ago by Zhc4zzz
  1

• It seems that the training memory is not reduced

  #29 opened a year ago by casillas1111
  1

• About training time.

  #23 opened a year ago by leexinhao
  2

• About CLIP-based adapter

  #28 opened a year ago by kaiw7
  1

• Training effiency

  #27 opened a year ago by Fanye12
  2

• Question about the Custom loss function on Recognition (like ViT model)

  #26 opened 2 years ago by andrewwang0612
  1

• something questions about hyper parameters and experiment results

  #24 opened 2 years ago by 007invictus
  1

• mmaction

  #22 opened 2 years ago by dmmSJTU
  1

• swin2d_base.py and swin2d_adapter_base.py

  #18 opened 2 years ago by XiudingCai
  1

• why the second dimension is n*b?

  #21 opened 2 years ago by 007invictus
  4

• Did you modify any config when change the frame_len from 8 to 16 or 32?

  #20 opened 2 years ago by leexinhao
  1

• Where is the function “init weights” called

  #17 opened 2 years ago by xuesongnie
  1

• T-MSA and S-MSA

  #16 opened 2 years ago by chaolistat
  3

• How can we freeze the layer?

  #14 opened 2 years ago by VoyageWang
  2

• SSv2 models

  #15 opened 2 years ago by or7r
  2

• Could you please share the swin transformer model weights on Kinetics 400?

  #13 opened 2 years ago by K0145vin
  2

• Maybe you should set `find_unused_parameters=True` in you config file?

  #11 opened 2 years ago by leexinhao
  4

• Can't reproduce Diving48 training

  #5 opened 2 years ago by ahnGeo
  1

• where you load the checkpoint of ViT backbone?

  #12 opened 2 years ago by SHIBOYA
  2

• where is your model?

  #10 opened 2 years ago by SHIBOYA
  1

• Visualization Experiments

  #9 opened 2 years ago by laohuijiadezhu
  1

• Training log available?

  #7 opened 2 years ago by simonJJJ
  5

• 在ViT_CLIP 中没有发现冻结层不做训练的代码，请问您这边在哪里做了处理，就是冻结预训练好的bcakbond权重

  #8 opened 2 years ago by xiaoyaod
  2

• About input frames and sampling interval

  #6 opened 2 years ago by BinhuiXie
  1

• Do you have pretrained model weights like ViT-S/16 or ViT-S/16？

  #4 opened 2 years ago by RayTang88
  1

• Do you have pretrained model weights like ViT-S/16 or ViT-S/16？

  #3 opened 2 years ago by RayTang88
  0

• Enquiry on the accessibility of SSV2 configs

  #2 opened 2 years ago by PeiqinZhuang
  1

• requirements.txt natplotlib => matplotlib?

  #1 opened 2 years ago by HolyDragon57
  1