You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
We want to clarify that this is an inference-only demo, that only supports our trained styles (denoted as 102, 103, 106, .etc, and the corresponding style images are presented in the examples). As can be seen, the selected style is None, which means this is an inference with the vanilla MUSE-Pytorch model, without our trained StyleDrop. We will update the gradio demo to better clarify its usage.
The gradio demo does not yet support training & inference with custom images, but you can still refer to the instructions here in our readme to train your custom StyleDrop weight.
Welcome further feedback on training your custom StyleDrop weights!
Could you provide information about the specifications of your local machine? Our code has been tested on Ubuntu 22.04 with a GPU that has 24GB of memory.
Not very clear, u can use fp16 clip by set prompt_model,_,_ = open_clip.create_model_and_transforms('ViT-bigG-14', 'laion2b_s39b_b160k',device='cpu')here. Have a try
Activity
encounter1997 commentedon Jul 7, 2023
Hi, thanks for your interest.
We want to clarify that this is an inference-only demo, that only supports our trained styles (denoted as 102, 103, 106, .etc, and the corresponding style images are presented in the examples). As can be seen, the selected style is None, which means this is an inference with the vanilla MUSE-Pytorch model, without our trained StyleDrop. We will update the gradio demo to better clarify its usage.
The gradio demo does not yet support training & inference with custom images, but you can still refer to the instructions here in our readme to train your custom StyleDrop weight.
Welcome further feedback on training your custom StyleDrop weights!
lucasjinreal commentedon Jul 7, 2023
@encounter1997 I see. When will a training support at least on my local machine?
encounter1997 commentedon Jul 7, 2023
Could you provide information about the specifications of your local machine? Our code has been tested on Ubuntu 22.04 with a GPU that has 24GB of memory.
lucasjinreal commentedon Jul 7, 2023
@encounter1997 Will 16GB mem work?
2blackbar commentedon Jul 7, 2023
how long it trains on 3090 , whats the max resolution?
zideliu commentedon Jul 8, 2023
If you set
config.sample_interval = Falsehere, 10 minutes enough, the max resolution is 256x256zideliu commentedon Jul 8, 2023
Not very clear, u can use
fp16clip by setprompt_model,_,_ = open_clip.create_model_and_transforms('ViT-bigG-14', 'laion2b_s39b_b160k',device='cpu')here. Have a try