search

caizhongang / SMPLer-X

Official Code for "SMPLer-X: Scaling Up Expressive Human Pose and Shape Estimation"
https://caizhongang.github.io/projects/SMPLer-X/
Other
1.01k stars 73 forks source link

[Confusion] Pretrained Model about ViTPose #63

Closed jameskuma closed 5 months ago

jameskuma commented 5 months ago

Dear authors, thank you for this awesome work!

I am working on this repo and I feel confused about ViTPose pretrained model.

As you describe in README.md, I should follow OSX to prepare models but I only see that OSX need me to download model with these names, like osx_vit_b.pth and etc.

However, I find in your file tree, the name of vit-pose model should be vitpose_base.pth and etc.

So I am confused about this, and could you help me clear this or you could share the link of these pretrained vitpose model that you use in your paper? (Since I also check vit-pose repo to find a bunch of weights and dont know which to download)

Best regards

awfuact commented 4 months ago

Hi @jameskuma, I am also confused about the ViTPose pretrained model. Which checkpoint did you end up downloading?

Dratlan commented 1 month ago

Hi @awfuact ,I'm also encounter this problem,did you figure it out? Thanks

awfuact commented 1 month ago

Hi @Dratlan, I found all pretrained things and models here: https://huggingface.co/camenduru/SMPLer-X/tree/main

Dratlan commented 1 month ago

@awfuact Thanks