eval EMDB

[xiaoyudanaa]

Exciting work. I have a problem with the error in the picture when I evaluate on EMDB(1). Do you know how to solve it? (Both 3DPW and RICH are runnable)

[yohanshin]

Hi @xiaoyudanaa , as mentioned at this line in the evaluation script, the current code requires EMDB to be downloaded and stored at ./dataset/. Please visit the official website to register and download the EMDB dataset to run the current evaluation. Sorry for the confusion!

[zehongs]

Hi @yohanshin , would you be willing to upload the implementation of RTE, ROE, and ERVE for testing the EMDB-2?

[JoeLeelyf]

Hi @yohanshin @xiaoyudanaa, how did you successfully evaluate WHAM on RICH? I try hard to make sure that I have strictly followed the setup instructions provided in the README file, and indeed successfully evaluate WHAM on 3DPW But when it comes to RICH, I encountered two main problems as follows:

1. 

2. The first one can be partly tackled by setting cfg.FLIP_EVAL = False, but I want to know what this parameter means exactly. The second may be due to the miss of smplx models, and I have noticed that you haven't mentioned this in your README file :)

[GaigeY]

Hi @yohanshin @xiaoyudanaa, how did you successfully evaluate WHAM on RICH? I try hard to make sure that I have strictly followed the setup instructions provided in the README file, and indeed successfully evaluate WHAM on 3DPW But when it comes to RICH, I encountered two main problems as follows: 1. 2. The first one can be partly tackled by setting cfg.FLIP_EVAL = False, but I want to know what this parameter means exactly. The second may be due to the miss of smplx models, and I have noticed that you haven't mentioned this in your README file :)

Hi, thanks for you experience sharing. To the second problem, you should download the THREE SMPLX_GENDER.npz files from the link provided by smplx, and put them into dataset/body_models/smplx/ folder.

[OliverSTH]

Exciting work. I have a problem with the error in the picture when I evaluate on EMDB(1). Do you know how to solve it? (Both 3DPW and RICH are runnable)

Hi, @xiaoyudanaa I would like to ask you why I can't evaluate Rich dataset when I run this command.

python -m lib.eval.evaluate_rich --cfg configs/yamls/demo.yaml TRAIN.CHECKPOINT checkpoints/wham_vit_w_3dpw.pth.tar

Do you do the same thing? I have the following problem. Do you know how to solve it?

Traceback (most recent call last):
  File "/home/dvlab/anaconda3/envs/wham/lib/python3.9/runpy.py", line 197, in _run_module_as_main
    return _run_code(code, main_globals, None,
  File "/home/dvlab/anaconda3/envs/wham/lib/python3.9/runpy.py", line 87, in _run_code
    exec(code, run_globals)
  File "/home/dvlab/WHAM/lib/eval/evaluate_rich.py", line 156, in <module>
    main(cfg, args)
  File "/home/dvlab/anaconda3/envs/wham/lib/python3.9/site-packages/torch/autograd/grad_mode.py", line 27, in decorate_context
    return func(*args, **kwargs)
  File "/home/dvlab/WHAM/lib/eval/evaluate_rich.py", line 67, in main
    flipped_batch = eval_loader.dataset.load_data(i, True)
  File "/home/dvlab/WHAM/lib/data/datasets/dataset_eval.py", line 31, in load_data
    target = self.__getitem__(index)
  File "/home/dvlab/WHAM/lib/data/datasets/dataset_eval.py", line 39, in __getitem__
    target = self.get_data(index)
  File "/home/dvlab/WHAM/lib/data/datasets/dataset_eval.py", line 110, in get_data
    target = self.prepare_inputs(index, target)
  File "/home/dvlab/WHAM/lib/data/datasets/dataset_eval.py", line 76, in prepare_inputs
    data = self.labels[self.prefix + key][index][1:]
IndexError: list index out of range

Thank you for your reply.