aipixel/AEMatter

Issues

• main分支复现

  #29 opened 5 months ago by caihaunqai
  7

• Can you provide a multi-machine training version of pytorch2.0?

  #26 opened 5 months ago by caihaunqai
  1

• Pytorch2.0版本是否不需要加载swin的权重？

  #27 opened 5 months ago by caihaunqai
  1

• 复现结果异常

  #28 opened 5 months ago by caihaunqai
  1

• 请教一下微调的内容

  #24 opened 9 months ago by hsdjkfnsfc
  7

• 如果不使用作者提供的预训练参数，而是自行训练，可以商用吗？

  #25 opened 8 months ago by PeideChi
  2

• About the inference result

  #22 opened 9 months ago by hsdjkfnsfc
  8

• AIM500

  #23 opened 9 months ago by linyiheng123
  1

• 请教细节 关于 patch size的定义

  #21 opened 9 months ago by astro-fits
  1

• Training sets do not have folder presentation instructions

  #17 opened 10 months ago by yuqilol
  1

• More details about RWA model

  #20 opened 10 months ago by peiwang062
  1

• update

  #18 opened 10 months ago by astro-fits
  1

• Bug? redundant contour line

  #19 opened 10 months ago by wangjia184
  2

• how many days are needed to reproduce the results

  #16 opened 10 months ago by ChenyiZhang007
  6

• 关于网络输出前景物体的图像

  #12 opened 10 months ago by astro-fits
  2

• 请问训练集 仅仅使用了 Compositiong-1K 吗

  #15 opened 10 months ago by astro-fits
  1

• Training dataset of Swin-Tiny

  #14 opened 10 months ago by viktorpovazhuk
  2

• train code release

  #4 opened 10 months ago by JWSunny
  4

• speed

  #7 opened 10 months ago by lingcong-k
  2

• 寻求 demo image

  #13 opened 10 months ago by NTUYWANG103
  1

• 训练学习率设置

  #11 opened a year ago by astro-fits
  2

• Are AEMatter and AEMatter+TTA the same weight model?

  #8 opened a year ago by shiwanlin
  1

• some great men can share some url's with weight, which can be easily use in great china?

  #10 opened a year ago by aiXia121
  0

• axis attention机制

  #9 opened a year ago by astro-fits
  7

• composition loss

  #6 opened a year ago by astro-fits
  2

• a question

  #5 opened 2 years ago by astro-fits
  2

• Possible issue with Multihead attention inputs

  #3 opened 2 years ago by VolodymyrAhafonov
  4

• Model outputs NaNs with Pytorch >= 2.0.1

  #2 opened 2 years ago by jacobbieker
  5