MCG-NJU/VideoMAE

Issues

• Overfitting in VideoMAE Model Fine-Tuning for Binary Classification on Home Camera Footage

  #129 opened 2 months ago by tgcandido
  0

• ckpt链接打开出错

  #128 opened 2 months ago by DragonWang-cell
  0

• TypeError: __init__() got an unexpected keyword argument 'all_frames'

  #77 opened 2 years ago by rahulstupa
  1

• length of video

  #127 opened 3 months ago by kimsekeun
  0

• TypeError: VisionTransformer.__init__() got an unexpected keyword argument 'pretrained_cfg'

  #126 opened 5 months ago by FVT34U
  1

• VideoMAE might not be a useful model for video reconstruction? Or perhaps it only learns the most generic distribution within patches?

  #120 opened 9 months ago by apptcom1123
  2

• KeyError: 'model'

  #119 opened 9 months ago by fengjingchehu
  2

• ucf101,hmdb51

  #122 opened 8 months ago by kffeng
  1

• 关于微调损失函数

  #124 opened 7 months ago by D-W-Y
  1

• How good model handle the different duration of clips?

  #116 opened 6 months ago by shinbehavior
  0

• How to conduct zero-shot evaluations?

  #125 opened 7 months ago by XuecWu
  0

• How to convert .pth to .bin

  #118 opened 9 months ago by fengzi456258
  1

• After obtaining the .pth file from training, how do I convert it into a .bin file for performing inference?

  #123 opened 7 months ago by caojiehui
  0

• Issue Encountered When Loading the Model: "pretrain_videomae_base_patch16_224"

  #105 opened 2 years ago by bbbdbbb
  5

• How to use VideoMAE for video regression task?

  #121 opened 9 months ago by YuHoChau
  0

• dist_init_required

  #90 opened 2 years ago by Malitha123
  1

• TODO/videomae_pretrain_base_patch16_224_frame_16x4_tube_mask_ratio_0.9_e1600/checkpoint-1599.pth

  #110 opened a year ago by xiaoli4881
  1

• How to obtain the reconstructed image for inference and masked

  #115 opened a year ago by hzxie99
  0

• UCF101

  #98 opened 2 years ago by wjj-w
  4

• AttributeError: module 'torch._C' has no attribute '_get_privateuse1_backend_name'

  #114 opened a year ago by abhisheksushil2003
  0

• the numberi of rebuild images is too small

  #113 opened a year ago by TuuSiwei
  0

• Rebuild Video

  #112 opened a year ago by TuuSiwei
  0

• HMDB checkpoint

  #111 opened a year ago by azabelo
  0

• VideoMAE ViT-H pre-train does not contain the decoder weights

  #89 opened 2 years ago by sandstorm12
  2

• The GPU memory usage of UCF and Kinetics is different.

  #109 opened a year ago by Backdrop9019
  0

• video xxx not correctly load during training

  #108 opened a year ago by JinChow
  0

• License of Kinetics-400

  #107 opened a year ago by joaopaulq
  0

• ModuleNotFoundError: No module named 'petrel_client'

  #106 opened a year ago by Kaicheng-Yang0828
  0

• the acc of small batch datasets is too low

  #84 opened 2 years ago by binbinjiang0505
  4

• About the encoder layer output

  #104 opened 2 years ago by Shar-01
  0

• The Batch size and training epoch not metch with paper

  #103 opened 2 years ago by Sumutan
  0

• Can I fine-tine it on a video dataset of 32 frames?

  #102 opened 2 years ago by Ha0Tang
  0

• What's the finetuning differences between ViT-B 80%acc and 81%acc?

  #100 opened 2 years ago by Vickeyhw
  0

• Questions about performence on ssv2

  #99 opened 2 years ago by wnzhyee
  0

• Fail to finetune from the provided pretrained model checkpoint on UCF101

  #92 opened 2 years ago by Yisen-Feng
  5

• BUG: Incorrect temporal indexing?

  #97 opened 2 years ago by rosenfeldamir
  0

• could you please provide me the weight of VideoMAE pre-trained on Kinetics-400,I want to use the the weight to extract the features of the thumos14

  #95 opened 2 years ago by Value-Jack
  4

• The dataset files in the link are not available

  #96 opened 2 years ago by Sumutan
  2

• How many videos are in your validation set?

  #94 opened 2 years ago by Sumutan
  3

• About pre-trained models

  #91 opened 2 years ago by 972821054
  0

• ViT-S and ViT-H models on huggingface

  #88 opened 2 years ago by sandstorm12
  0

• MoCoV3 Training Configuration

  #87 opened 2 years ago by fmthoker
  0

• Can VideoMAE be used to learn the motion characteristics and appearance characteristics of objects in videos?

  #86 opened 2 years ago by summersnowfish
  0

• No such file or directory: '/home/zhiyuan/img_diff_sthv1_train.json'

  #79 opened 2 years ago by rouge012
  1

• learnable position embedding

  #83 opened 2 years ago by kun-dragon
  1

• ssv2 dataset acc

  #75 opened 2 years ago by an1018
  2

• dataloader hinder the training speed

  #81 opened 2 years ago by valencebond
  1

• VIT-S initilization

  #80 opened 2 years ago by G-JWLee
  1

• why do not you use [CLS] token?

  #78 opened 2 years ago by LinB203
  0

• Can't download SSv2 ViT-B pre-trained model

  #74 opened 2 years ago by sunilhoho
  1