T. Xu comments

Results 30 comments of


                                            T. Xu

AttributeError: 'EMA' object has no attribute '_optimizer_state_dict_pre_hooks'

including ```python self._optimizer_state_dict_pre_hooks = OrderedDict() self._optimizer_state_dict_post_hooks = OrderedDict() ``` in __init__ method of EMA class work for me @RiceRollsMan @qiangwang57 @codgodtao

The question about different size images to compress

Currently the image size is determined by your base diffusion model. This repo use pretrained ADM with image size 256^2. So we can only work with 256^2 image for one...

the result that imagenet model generated is not good enough, I think the diffusion config sets more steps?

Hi @AlbertLin0, this result is pretty strange and we have never seen such a reconstruction. We believe it is a bug, can you provide the exact command line and config?

the result that imagenet model generated is not good enough, I think the diffusion config sets more steps?

> hi @tongdaxu, The command line is: > > ``` > --model_config=configs/imagenet_model_config.yaml > --diffusion_config=configs/diffusion_config.yaml > --task_config=configs/elic_config.yaml > --save_dir=results > ``` > > the config: I just change the 'q' from...

the result that imagenet model generated is not good enough, I think the diffusion config sets more steps?

> > > hi @tongdaxu, The command line is: > > > ``` > > > --model_config=configs/imagenet_model_config.yaml > > > --diffusion_config=configs/diffusion_config.yaml > > > --task_config=configs/elic_config.yaml > > > --save_dir=results >...

the result that imagenet model generated is not good enough, I think the diffusion config sets more steps?

> > > > > hi @tongdaxu, The command line is: > > > > > ``` > > > > > --model_config=configs/imagenet_model_config.yaml > > > > > --diffusion_config=configs/diffusion_config.yaml >...

the result that imagenet model generated is not good enough, I think the diffusion config sets more steps?

Hi @AlbertLin0, I have asked my intern @AuroraZengfh to run the ELIC Imagenet from scratch, and the result is fine. I have asked him to post the result and his...

Update on Diffusion model-based Image/Video Compression

> Thanks for your very excellent works, we will incorporate these works into our repo, and revise the corresponding parts. Thanks for including this, some bug fixs: * [Lossy Image...

val.log中的PSNR值时高时低？

Hi @One1209 @ranformer, This phenomena is known as perception-distortion trade-off. The output psnr is computed by E[X0|Xt, y]. In the first step, it is given by E[X0|XT, y]. Technically speaking,...