Awj2021

Results 17 comments of Awj2021

> ![image](https://user-images.githubusercontent.com/27964642/174237016-05e3aad2-08c6-4614-a6c2-5c374a67877a.png) ![image](https://user-images.githubusercontent.com/27964642/174237145-6465ef71-25c8-4b5c-abfd-4377fe68c540.png) Have you met this error? Hi, I met the same error, did you fix this problem? Thanks.

hi, @Arthur151, when I use this setting of temporal_optimize=true, and the both people, although the real distance is very large, these two person's meshes are overlaped together. Why this happened?...

hi, guys,have you solved this problem?

@mrapplexz @dribnet hi, Thank you for your amazing ideas, but there are some points confused me. When resuming the model, how to set the training steps? e.g., , I have...

And I have another question as showed [issues/93](https://github.com/CompVis/taming-transformers/issues/93), If use different dataset (e.g., medical Image dataset) to finetune the method, the parameter `disc_start = 0` showed in [https://heibox.uni-heidelberg.de/d/8088892a516d4e3baf92/](https://heibox.uni-heidelberg.de/d/8088892a516d4e3baf92/) maybe not...

> here Yeah, I could download these .npy files

> In addition, you can generate the .npy for category name embds by `tools_draft/construct_CatName2vec.py` file: tools_draft/construct_CatName2vec.py is used for generating the vidvrd_PredNameEmb.npy and vidvrd_EntiNameEmb.npy, but how about the pred_bias_matrix_vidvrd.npy? Thank...

> > > here > > > > > > Yeah, I could download these .npy files > > Could you please tell how you did it? I directly download...

have you solved this problem? @Zhongxyyyy @2021wangkai