robvanvolt

> Why can't the most important tests be run on CPU ? > > Even if it's takes 5min to do 3 steps of batch size 2 on cpu, it's...

> @afiaka87 would you mind sending an updated discord link? the current one is broken afaict - thanks! https://discord.gg/gVuU2YE5hd

Maybe we should convert our datasets to tar.gz and work with WebDatasetloaders? This might be a faster approach than optimizing reading text and image folders, doesn't it? But cool anyway...

Added early beta support for WebDatasets: https://github.com/lucidrains/DALLE-pytorch/pull/280

> Added early beta support for WebDatasets: > #280 Added full support for WebDatasets: https://github.com/lucidrains/DALLE-pytorch/pull/280 Does anyone want to try the new feature out or review the changes?

The link in this repo's readme github is not valid anymore, here is a new one that does not expire anymore: https://discord.gg/gVuU2YE5hd @lucidrains could you update that link on the...

Oh yeah, I think it was a special invitation link generated by booster level 3.. Here is a new invitation link, that should never expire... :-) https://discord.gg/xBPBXfcFHd

Doen't this link from the readme work anymore?:) This link works for me: https://discord.gg/xBPBXfcFHd

> The CUDA memory allocation values don't make sense to me. > > > `Tried to allocate 1.93 GiB (GPU 0; 23.70 GiB total capacity; 15.29 GiB already allocated; 1.54...

In theory, the model "might" miss some features, but in the CogView paper the authors state that both sparse and full attention yield similar results for image generation - only...