search

lucidrains / DALLE2-pytorch

Implementation of DALL-E 2, OpenAI's updated text-to-image synthesis neural network, in Pytorch
MIT License
11.03k stars 1.07k forks source link

Noise Output Result Issue #61

Closed goldiusleonard closed 2 years ago

goldiusleonard commented 2 years ago

@lucidrains I have the same issue with #54 . I have copied the training method in the readme, but I still got the noise output result. Can you see my code and check what is wrong with it?

https://github.com/goldiusleonard/Dalle2_pytorch_project

lucidrains commented 2 years ago

@goldiusleonard yea, so the training loop is actually more involved, but i hope to automate that all away in a CLI tool, by end of the month. it'll end up exactly like https://github.com/lucidrains/lightweight-gan no-nonsense

goldiusleonard commented 2 years ago

@goldiusleonard yea, so the training loop is actually more involved, but i hope to automate that all away in a CLI tool, by end of the month. it'll end up exactly like https://github.com/lucidrains/lightweight-gan no-nonsense

so any solution for my issue? Or I just need to train it on more epochs?

lucidrains commented 2 years ago

@goldiusleonard once you see the (wip) tag removed from the Training section, then it is ready for you

right now it is only accessible to deep learning practitioners

lucidrains commented 2 years ago

@goldiusleonard if you wish to collaborate with people who work with artificial intelligence, try joining the discord and be a fly on the wall. you'll learn lots

lucidrains commented 2 years ago

@goldiusleonard depending on level of sponsorship, i should be able to carry the project all the way to being command-line trainable by anyone, even your grandparents :laughing: