Glory Chen

Problems of IWAE ELBO Loss

5

Hi Anand and all, As weighting of samples, `weight` should be detached from the current computational graph for the expected optimization objective, right? See https://github.com/AntixK/PyTorch-VAE/blob/8700d245a9735640dda458db4cf40708caf2e77f/models/iwae.py#L155

Modification for Test

3

It is needed to add a line `self.save_hyperparameters()` in https://github.com/AntixK/PyTorch-VAE/blob/8700d245a9735640dda458db4cf40708caf2e77f/experiment.py#L16 initialization to save hyperparameters (i.e., `vae_model` and `params` here), for successfully calling `LightningModule.load_from_checkpoint(PATH)` and `runner.test()` afterwards.

GPU Mem Double

1

Hi Shitong, A great job! May I ask why GPU mem double after some iterations? E.g., from 8G to 14G. Thanks,

Setup Problems: Recent Version of TF Probability and Dlerror

4

Hi, Thanks for your excellent work and released code repo. Following https://github.com/yang-song/score_sde_pytorch#dependencies, encountered the problem, ```text 2021-06-23 20:51:42.320605: W tensorflow/stream_executor/platform/default/dso_loader.cc:60] Could not load dynamic library 'libcudart.so.11.0'; dlerror: libcudart.so.11.0: cannot open...

Unusual to Use Large Filters

1

Hi Partha @ParthaEth , Thanks for the code. As referred to, https://github.com/ParthaEth/Regularized_autoencoders-RAE-/blob/9478b8f781f7229807a0d7c4ea92a7c9c7994bfa/models/rae/rae_celeba.py#L80 it seems unusual to me to use such large filters in Conv generative models? Why do that? I...

https://github.com/hhb072/IntroVAE/blob/c8ce5d291fe8e66189d70b3ceddc2eb2266d3742/main.py#L191 It seems more natural of the converse case, i.e., relu(...).mean().

Not Implementing IWAE?

2

Hi, Thanks for your good work and repo. May I ask did you implement the exact IWAE as a tighter lower bound, in case I missed it? See the reference:...

MSCOCO Background

3

Hi @mks0601 and folks, May I ask where I can find "images of MSCOCO dataset that do not include humans to augment the backgrounds following Moon et al. [30]" to...

Hi @YuliangXiu @Jeff-sjtu , May I ask where the below is used? Thx & regards, https://github.com/YuliangXiu/PoseFlow/blob/4b2f266bbea99790904ebf0dc6c54ef25809ab5f/utils.py#L81

Hi @nkolot @geopavlakos et al., Could you help release H36M (OP) used in Tab. 3 to reproduce the results? Thx & best,