ik_vision

@mkocabas this is a major progress on VIBE, thank you for the transparency and your open-science approach. A minor technical remark for reproducibility, comparing **best** accuracy of the experiments is...

using demo.py command line argument: `--save_obj: Save output meshes as .obj files. ` The obj output file include the 3d mesh in textual format. To extract the mesh the prefix...

@durgesh17media you are right, sorry, this flag exists only in VIBE https://github.com/mkocabas/VIBE/blob/master/demo.py#L378-L379 If you want to extract the mesh from SPIN, vertices are stored in pred_vertices and the faces in...

Thank you for the quick reply, I fixed my solution to: `train_data=imdb.images.data(:,:,imdb.images.set==1);` `imageMean =mean(train_data(:))` It made Less than a half a percent diff in the imageMean value

This may help, by default the checkpoint is loaded into GPU 0 https://github.com/nkolot/SPIN/blob/b95a00a7c0147f2c5bee0874ba0972c6389b6f99/demo.py#L105 Try changing it into: `gpu_num = 1` `device = torch.device('cuda',gpu_num)` `checkpoint = torch.load(args.checkpoint, map_location=device)`

I think this paper tackles the problem of 13 body measurements: https://hal.inria.fr/hal-02162159/file/virtual_caliper_camera_ready.pdf and another relevant paper is: http://files.is.tue.mpg.de/black/papers/wacv14_measprediction.pdf 

@harshitv804 I am working on extending chronos to add covariates using an lgbm regression head on top of univariate embeddings If you want to assist me to progress on this...

@hsm207 Do you want to forecast all 3 variables future based on all 3 variables past value? Did you try https://huggingface.co/Salesforce/moirai-1.0-R-large?