raoyongming/DenseCLIP

Issues

• Question about little FPN in ViT-B

  #14 opened 3 years ago by Unrealluver
  2

• shape related question

  #55 opened 10 months ago by SKevin673
  0

• A little question about dimensions

  #54 opened a year ago by lxr-1204
  0

• some question about pixel-text matching loss

  #53 opened a year ago by Yu-zhengbo
  1

• CUDA out of memory

  #52 opened a year ago by MrChenNX
  1

• Training details

  #51 opened a year ago by rezaqorbani
  0

• Request about the pre-trained model of Swin Transformer+DenseCLIP

  #50 opened a year ago by tanbuzheng
  0

• super(SingleStageDetector, self).__init__(init_cfg) TypeError: __init__() takes 1 positional argument but 2 were given

  #49 opened a year ago by YangJae96
  1

• Is default training iterations enough to reach the paper performance?

  #30 opened a year ago by williamlus
  1

• question about how to use vit backbone in detection

  #43 opened a year ago by tandangzuoren
  1

• question on the device

  #44 opened a year ago by duoxiangqinzuo
  1

• downloading the pretrained weights

  #46 opened a year ago by elhamAm
  1

• Question about DenseCLIP for Any Visual Backbone

  #47 opened a year ago by needsee
  6

• Question about training process

  #48 opened a year ago by 0xac9527
  2

• Code for any backbone(ImageNet) experiments on ADE20K segmentation

  #45 opened a year ago by ByungJu-Woo
  0

• the results of DenseCLIP on Cityscapes

  #22 opened 3 years ago by Richardych
  3

• New MMCV and MMSegmentation version

  #36 opened 2 years ago by Jtao0818
  2

• code about Pre-model prompting

  #11 opened 3 years ago by Ahnsun
  3

• Code

  #42 opened 2 years ago by kexibudongshi
  2

• vit-b-denseclip for semantic segmentation is lost

  #41 opened 2 years ago by sunwhw
  1

• about single-scale and multi-scale settings

  #38 opened 2 years ago by chenhao-zju
  1

• What does the different contexts_length setting based on? What is the meaning of separation?

  #40 opened 2 years ago by GanPeixin
  4

• Loading pretrained CLIP parameters but tuncate context_length in positional_embedding?

  #39 opened 2 years ago by laisimiao
  2

• An error while saving the model in ONNX format

  #32 opened 2 years ago by Virgilzzz
  1

• Questions about text input

  #35 opened 2 years ago by Virgilzzz
  2

• What does "We fix the text encoder during training" mean? Does it mean that the parameters are not updated during training?

  #33 opened 2 years ago by sunwhw
  3

• Prompt learning via CoOp

  #34 opened 2 years ago by anurag-198
  2

• About prompt text

  #31 opened 2 years ago by good-demo
  3

• Question about implementation of CLIPResNet

  #29 opened 2 years ago by littlepenguin89106
  2

• Open set inference without training?

  #25 opened 2 years ago by Colin97
  6

• [critical bug] The text encoder is also updated.

  #28 opened 2 years ago by SeongwoongCho
  2

• Misaligned params

  #27 opened 2 years ago by sfchen94
  2

• A stupid question about auxiliary loss for Object detection & instance segmentation.

  #26 opened 2 years ago by waxnkw
  4

• Single GPU error

  #24 opened 2 years ago by Virgilzz
  3

• Questions about the architecture

  #21 opened 3 years ago by xiaoachen98
  2

• Question about ADE20K dataset

  #23 opened 3 years ago by wenyuqing
  2

• Question about CLIPVisionTransformer

  #20 opened 3 years ago by aniki-ly
  2

• Can not reproduce the result of DenseCLIP-R50.

  #18 opened 3 years ago by Richardych
  8

• Question about inference setting

  #17 opened 3 years ago by dneirfi
  3

• dimension error when load ViT-B weight

  #19 opened 3 years ago by xuzhang1199
  1

• what is the value of gamma?

  #16 opened 3 years ago by Richardych
  1

• Questions about the details of configuration of RN50-CLIP

  #13 opened 3 years ago by JasonLin1998
  4

• Where can I download this （RN102.pt）?

  #12 opened 3 years ago by usr922
  2

• dim unsigned

  #10 opened 3 years ago by Ahnsun
  2

• multi-gpu error

  #9 opened 3 years ago by eternaldolphin
  3

• Query on Inference Setting

  #8 opened 3 years ago by sauradip
  3

• question about eos_indx in model.py

  #4 opened 3 years ago by qiulesun
  11

• question about any backbone experiments on ADE20K segmentation

  #7 opened 3 years ago by wanglixilinx
  5

• Some questions of ViT-B-DenseCLIP

  #6 opened 3 years ago by lixiangMindSpore
  6

• ADE20K batchsize

  #5 opened 3 years ago by RainHxj
  6