train error about h36m

[SaterZhan]

hello，i meet a question like this，i failed to resolve it ，need your help，thanks
INFO:datasets.base_dataset:len of h36m-p2-mosh: 26859
0%| | 0/200 [00:23<?, ?it/s]
Traceback (most recent call last):_Jan05-08-53-12-BCm Epoch 0: 0%| | 1/4877 [00:12<16:54:58, 12.49s/it]
File "train.py", line 69, in
main(None, ngpus_per_node, options)
File "train.py", line 38, in main
trainer.fit()
File "/home/data1/zxt/PyMAF-master/core/trainer.py", line 502, in fit
self.train(epoch)
File "/home/data1/zxt/PyMAF-master/core/trainer.py", line 278, in train
out = self.train_step(batch)
File "/home/data1/zxt/PyMAF-master/core/trainer.py", line 324, in train_step
gt_out = self.smpl(betas=gt_betas, body_pose=gt_pose[:,3:], global_orient=gt_pose[:,:3])
File "/home/data2/anaconda1/envs/pymaf/lib/python3.6/site-packages/torch/nn/modules/module.py", line 493, in call
result = self.forward(*input, **kwargs)
File "/home/data1/zxt/PyMAF-master/models/smpl.py", line 34, in forward
smpl_output = super().forward(*args, **kwargs)
File "/home/data2/anaconda1/envs/pymaf/lib/python3.6/site-packages/smplx/body_models.py", line 376, in forward
self.lbs_weights, pose2rot=pose2rot, dtype=self.dtype)
File "/home/data2/anaconda1/envs/pymaf/lib/python3.6/site-packages/smplx/lbs.py", line 179, in lbs
v_shaped = v_template + blend_shapes(betas, shapedirs)
File "/home/data2/anaconda1/envs/pymaf/lib/python3.6/site-packages/smplx/lbs.py", line 265, in blend_shapes
blend_shape = torch.einsum('bl,mkl->bmk', [betas, shape_disps])
File "/home/data2/anaconda1/envs/pymaf/lib/python3.6/site-packages/torch/functional.py", line 211, in einsum
return torch._C._VariableFunctions.einsum(equation, operands)
RuntimeError: size of dimension does not match previous size, operand 1, dim 2
pymaf_res50_as_lp3_mlp256-128-64-5_Jan05-08-53-12-BCm Epoch 0: 0%|

[HongwenZhang]

Hi, this issue may be raised by the incompatible number of the shape components in SMPL (I am not so sure). We use 10 shape components with version 0.1.13 of smplx. Please also refer to issue #4.

[SaterZhan]

thanks so much, you are right! i'm wrong as used another version of SMPL!