zyxElsa/InST

Issues

• Questions about the training time

  #55 opened 6 months ago by Meimeiainaonao
  1

• 'openai/clip-vit-large-patch14'

  #52 opened 7 months ago by yang12369
  2

• InST.ipynb doesn't match the original method, it's not the stylisation process in the original paper.

  #56 opened 5 months ago by BBABM
  0

• Fine-tuning broken

  #54 opened 6 months ago by domef
  1

• !!Help!!InST打不开style_image路径

  #22 opened 6 months ago by neverbiasu
  7

• Unable to reproduce paper results，无法复现论文结果

  #33 opened a year ago by xddun
  6

• what version of pytorch-lightning works best with main.py?

  #53 opened 7 months ago by ezralynch
  0

• a lot of missing keys while loading the FrozenCLIPEmbedder checkpoint

  #51 opened 8 months ago by musetee
  0

• How to prepare the Personalized dataset?

  #11 opened a year ago by kravrolens
  2

• Can I use several images for training?

  #30 opened a year ago by carvychen
  2

• Style doesn't transfer at all when running InST.ipynb with empty prompt.

  #42 opened a year ago by heorhiiberezovskyi
  7

• train

  #45 opened a year ago by OrangeZz0331
  3

• bug in embedding_manager

  #50 opened 8 months ago by nancy6o6
  1

• How about your training dataset?

  #6 opened 8 months ago by BossunWang
  5

• Torch._six

  #49 opened 9 months ago by arrrrr3186
  0

• Trained results are not so good

  #15 opened 2 years ago by sunkyya
  2

• the results are bad

  #48 opened 10 months ago by rshape
  0

• NameError: name 'trainer' is not defined

  #44 opened a year ago by happypuppy128
  9

• someting seems to be missing in main.py

  #47 opened 10 months ago by OrangeZz0331
  0

• How to train on multiple GPUs

  #46 opened 10 months ago by zhuangtu1203
  0

• trainer.fit(model, data) in main.py

  #13 opened 2 years ago by ssukii
  8

• Inconsistencies between my test results and those of the paper

  #43 opened a year ago by hhhhhh45
  5

• What license is this under?

  #10 opened a year ago by Njasa2k
  0

• Can a cartoon or anime image be transferred back to a real image?

  #41 opened a year ago by D-Mad
  0

• Question about why the predict noise is used ?

  #35 opened a year ago by Ed-ivan
  1

• Input type (torch.cuda.FloatTensor) and weight type (torch.FloatTensor) should be the samethis may be cased by the code in line 1112 in ddpm.py, plz change the code like this, but im not sure, plz let me see the hole issue report

  #36 opened a year ago by codemakerzz
  2

• Load pre-trained model:CLIP transformer encoder

  #38 opened a year ago by m-violet-s
  2

• During the training process, I encountered the following issue. Has anyone encountered a similar problem? Please help me to solve it, thank you!

  #39 opened a year ago by lzlzlz1999
  2

• 在运行inST.py时出现了Process finished with exit code 139 (interrupted by signal 11: SIGSEGV)

  #40 opened a year ago by juewangdelty
  1

• No "models/sd/sd-v1-4.ckpt"

  #37 opened a year ago by hhhhhh45
  0

• training time for "main.py"

  #26 opened a year ago by fikry102
  2

• you will release yours embeddings.pt model?

  #9 opened a year ago by zhanghongyong123456
  3

• RuntimeError: shape '[4, 1, 768]' is invalid for input of size 768

  #17 opened a year ago by pokameng
  1

• using prompt

  #34 opened a year ago by noakraicer
  0

• Stochastic Inversion and embedding.pt

  #27 opened a year ago by Gynjn
  1

• 8张16G的卡，也报CUDA out of memory，得多大的显存啊

  #31 opened a year ago by jiafengshen
  1

• 训练结果较差

  #29 opened a year ago by tro0o
  4

• When training the Diffusion model（main.py）, two processes are found to be open. What is the reason

  #32 opened a year ago by jiafengshen
  0

• Request for exemplar images

  #28 opened a year ago by emjay73
  0

• misunderstanding

  #24 opened a year ago by fikry102
  0

• finished

  #25 opened a year ago by fikry102
  0

• BUG

  #12 opened 2 years ago by ssukii
  2

• Different style_dir output the same result

  #23 opened a year ago by answerBoom
  0

• Training doesn't work

  #21 opened 2 years ago by luckyhzt
  0

• checkpoint.clear

  #19 opened 2 years ago by tjl1
  0

• How to separate style and content in the reference image?

  #20 opened 2 years ago by Sutongtong233
  0

• Is it possible that image x is not equal image y when training?

  #18 opened 2 years ago by zhangquanwei962
  0

• Dataset

  #16 opened 2 years ago by fangxy100
  0

• Can you run and test the project with a dedicated GPU of 10gb and a shared GPU of 32gb?

  #14 opened 2 years ago by 123456klk1
  0

• Clarification on style-transfer details

  #7 opened 2 years ago by zhihongp
  0