ermongroup/ddim issues and pull requests

#41 - Question in paper for accelerated sampling formula

Issue - State: open - Opened by LokiXun about 1 month ago

#41 - Question in paper for accelerated sampling formula

Issue - State: open - Opened by LokiXun about 1 month ago

#40 - sample seq problem

Issue - State: open - Opened by redcping about 1 month ago

#40 - sample seq problem

Issue - State: open - Opened by redcping about 1 month ago

#39 - Replace torch._six.string_classes by str

Pull Request - State: open - Opened by ghabault about 2 months ago

#39 - Replace torch._six.string_classes by str

Pull Request - State: open - Opened by ghabault about 2 months ago

#38 - Update celeba.py to avoid warnings and issue when downloading zip

Pull Request - State: closed - Opened by ghabault about 2 months ago

#38 - Update celeba.py to avoid warnings and issue when downloading zip

Pull Request - State: closed - Opened by ghabault about 2 months ago

#37 - Calculate FID with CelebA or LSUN datasets

Issue - State: open - Opened by parkjjoe 3 months ago

#37 - Calculate FID with CelebA or LSUN datasets

Issue - State: open - Opened by parkjjoe 3 months ago

#36 - FID evaluation terminated

Issue - State: closed - Opened by parkjjoe 4 months ago - 1 comment

#35 - Why use asymmetric padding in downsample?

Issue - State: open - Opened by tian-2024 5 months ago

#34 - implemented upscaling model

Pull Request - State: closed - Opened by vicliv 6 months ago

#33 - Loss is not going down

Issue - State: open - Opened by TerryMelody 6 months ago - 4 comments

#32 - Wrong fid in cifar10

Issue - State: open - Opened by HGGshiwo 8 months ago

#31 - without "train" function!!!!

Issue - State: open - Opened by hhhhhh45 8 months ago

#30 - Regarding the issue of excessive FID

Issue - State: closed - Opened by xzmy-Dreamleft 12 months ago - 2 comments

#29 - Could you please provide guidance on the calculation of FID?

Issue - State: closed - Opened by xzmy-Dreamleft about 1 year ago - 1 comment

#28 - Questions about noise distributions during training and sampling

Issue - State: closed - Opened by gitwangxf about 1 year ago - 1 comment

#27 - Generation Process not producing quality images.

Issue - State: closed - Opened by VishyTheFishy about 1 year ago - 2 comments

#26 - run sampling process with cifar10 dataset

Issue - State: closed - Opened by Sugar929 about 1 year ago - 1 comment

#25 - using DistributedDataParallel

Issue - State: closed - Opened by sostupidDog over 1 year ago - 2 comments

#24 - train loss goes very large

Issue - State: closed - Opened by imdoublecats over 1 year ago - 5 comments

#23 - why the activation functions formed like this in the code?

Issue - State: closed - Opened by ShiningorDying over 1 year ago - 1 comment

#22 - DDIM inversion

Issue - State: closed - Opened by Passenger12138 over 1 year ago - 4 comments

#21 - test data

Issue - State: open - Opened by tianzhijiaoziA over 1 year ago

#20 - Question about Lemma 1 in the paper

Issue - State: open - Opened by buttercutter over 1 year ago

#19 - How about the training setting of CelebA model

Issue - State: open - Opened by jsttlgdkycy almost 2 years ago - 3 comments

#18 - got an unexpected keyword argument `eta`

Issue - State: open - Opened by wyf0912 almost 2 years ago - 5 comments

#17 - FID of DDPM on CIFAR-10

Issue - State: closed - Opened by chenweixin107 almost 2 years ago - 3 comments

#16 - Transferability to transformers

Issue - State: closed - Opened by anthony-mendil almost 2 years ago - 2 comments

#15 - Does this repo support multi-GPUs?

Issue - State: closed - Opened by forever208 almost 2 years ago

#14 - Add link to Diffusers library

Pull Request - State: closed - Opened by patrickvonplaten almost 2 years ago

#13 - Add requirements.txt

Pull Request - State: closed - Opened by bell-one almost 2 years ago

#12 - HOW to use the pretrained model of celeba

Issue - State: open - Opened by pokameng almost 2 years ago

#11 - lsun.py can't load classes list

Issue - State: open - Opened by ryhhtn almost 2 years ago

#10 - sampling code of DDPM

Issue - State: closed - Opened by zzw-zwzhang about 2 years ago

#9 - reconstruction from latent code

Issue - State: closed - Opened by winnechan over 2 years ago - 7 comments

#8 - Question in the paper

Issue - State: closed - Opened by sunwoo76 over 2 years ago - 1 comment

#7 - got wrong FID using pretrained model

Issue - State: closed - Opened by ztyang196 over 2 years ago - 3 comments

#6 - Question in results

Issue - State: closed - Opened by zzw-zwzhang over 2 years ago - 1 comment

#5 - FID of CIFAR-10

Issue - State: closed - Opened by Mq-Zhang1 over 2 years ago

#4 - How to train a better model?

Issue - State: closed - Opened by ghost over 2 years ago - 1 comment

#3 - FID of DDPMs on CIFAR-10

Issue - State: closed - Opened by GloryyrolG almost 3 years ago - 3 comments

#2 - Why not clamp in generalized_steps?

Issue - State: closed - Opened by GloryyrolG almost 3 years ago - 1 comment

#1 - CelebA

Issue - State: closed - Opened by ghost about 3 years ago - 1 comment

GitHub / ermongroup/ddim issues and pull requests