Faster pytorch batched version.#11
Merged
CalciferZh merged 2 commits intoCalciferZh:masterfrom Feb 2, 2019
Merged
Conversation
I made this very fast pytorch implementation of your SMPL model for my work with the SMIL model. It took me all of monday, and then it turned out you had already made a pytorch version when I was done with my code. :) I believe this should work with both the SMPL and SMIL model. It is very very fast on my computer, only limited by the memory, and it works with sparse tensors too! (Saving a lot of said memory) I hope my code is not too messy, I don't have time to clean it up but I hope it is helpful to you. Try it out! Thank you so much for your work, it helped me a lot. SMIL: https://www.iosb.fraunhofer.de/servlet/is/82920/ If you have any questions feel free to ask. Best regards, Sebastian.
Now it is made possible to give None as model_path, which means it should not load the model. Then you can instead load from a state dict. This state dict must contain 'kintree_table', but otherwise partial loading is also possible. See https://pytorch.org/tutorials/beginner/saving_loading_models.html for examples.
Owner
|
Should we change the name of this repo? Since now we also support SMIL model. |
Contributor
|
@sebftw Thanks for your work! However I cannot find smil_np inside the repo, can you tell me how to run the test script? Thanks! |
Contributor
Author
|
You can replace it with smpl_np. 🙂 It is agnostic to which of the two models you use. |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
3 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
I made this very fast pytorch implementation of your SMPL model for my work with the SMIL model.
It took me all of monday, and then it turned out you had already made a pytorch version when I was done with my code. :)
I believe this should work with both the SMPL and SMIL model. It is very very fast on my computer, only limited by the memory, and it works with sparse tensors too! (Saving a lot of said memory)
I hope my code is not too messy, I don't have time to clean it up, but I hope it is helpful to you.
Try it out!
Thank you so much for your work, it helped me a lot.
SMIL: https://www.iosb.fraunhofer.de/servlet/is/82920/
If you have any questions feel free to ask.
Best regards, Sebastian.
PS: I gave up on making the model able to be loaded from a state dict. It turns out that you have to know the shape of the parameters/buffers for this to be possible.