While executing python3 fit.py --img data/000002.jpg, I receive the following error. The installed version of cuda on my system is 11. What should I do?

Traceback (most recent call last):
  File "fit.py", line 6, in <module>
    from core.models import ReconModel
  File "/home/tface/Projects/torch/3DMM-Fitting-Pytorch/core/models.py", line 5, in <module>
    from pytorch3d.renderer import (
  File "/home/tface/.virtualenvs/torch/lib/python3.6/site-packages/pytorch3d/renderer/__init__.py", line 3, in <module>
    from .blending import (
  File "/home/tface/.virtualenvs/torch/lib/python3.6/site-packages/pytorch3d/renderer/blending.py", line 9, in <module>
    from pytorch3d import _C
ImportError: libcudart.so.10.1: cannot open shared object file: No such file or directory


$ nvcc --version
nvcc: NVIDIA (R) Cuda compiler driver
Copyright (c) 2005-2020 NVIDIA Corporation
Built on Wed_Jul_22_19:09:09_PDT_2020
Cuda compilation tools, release 11.0, V11.0.221
Build cuda_11.0_bu.TC445_37.28845127_0

Thank you for sharing your work, please which parameters exactly correspond to obtain a good result ( preserved identity) and on what basis should I modify), because the output image does not look like the input (very smooth).

Now I make a group of exp basis data by myself. And I want to add it in extra expression basis.
Should I must use the same point sequence of BFM? If I must do this, how can I transform my own data?
Thx.

Thank you for your awesome work!
After running your project, I get the reconstructed face image and *.obj file, can you tell me how to get the rendered image just like these?

Hello,
I am trying to use my own statistical model with your code.
What is exactly is face_texture in BFM09ReconModel.forward ? I have dimensions incompatibility with a tensor of texture so I suspect it actually refers to vertex colors. correct ?

Also, what is BFM09ReconModel.point_buf ?

Thanks for your repo first.
I find the closed eye can not be fit.
Input :

Output:

The question is how to make the eyes close?

I try to use colab to run the code
I use the same data as readme.md
When i execute “!python convert_bfm09_data.py”，I get the following error:

Traceback (most recent call last):
File "convert_bfm09_data.py", line 102, in
transferBFM09()
File "convert_bfm09_data.py", line 38, in transferBFM09
expPC, expEV = LoadExpBasis()
File "convert_bfm09_data.py", line 15, in LoadExpBasis
expMU.fromfile(Expbin, 3*n_vertex)
ValueError: bytes length not a multiple of item size

Hello,

first of all thank you for the project.
I have an issue running your python script fit_single_img.py and Im getting this error :

RuntimeError: Expected all tensors to be on the same device, but found at least two devices, cuda:0 and cpu!
Are there any commands to specify the device?

Thank you in advance.

` vs = torch.matmul(vs, self.reverse_z.repeat((batchsize, 1, 1))) + self.camera_pos

aug_projection = torch.matmul(vs, self.p_mat.repeat((batchsize, 1, 1)).permute((0, 2, 1)))`
Hi, many thanks for your great work! I don’t know if I understand it right, the first line represents the orthographic projection? The second line indicates that the picture coordinate system is converted to the pixel coordinate system?

Here is your code

face_shape = torch.matmul(face_shape,reverse_z) + camera_pos
I don't know what it really means？

Hi I have an issue finding this dataset. I think the mat files are the 2009 versions and the new version doesn't seem to contain a matlab file at all

Hi, ascust, this repository is great. I am wondering how to prepare and train my own data if I want to not only reconstrct face but other components of a head？

Hi~

Thanks for this wonderful repo.
I notice that there is file named 'facemodel_info.mat', and we have to build the bfm model using this file. I'm wondering how to get this file. Especially I don't understand what 'point_buf' means. It seems 'point_buf' is used to compute the surface norm. But could you provide more information/explanation about this file? Thank you.

How to rotate and display the resultant 3D face as done on https://github.com/YadiraF/face3d.

Like this gif:https://github.com/ascust/3DMM-Fitting-Pytorch/blob/master/gifs/demo.gif

Hi, thanks for your work!
I am wonder why the function in BFM09ReconModel below?
face_shape = face_shape -
self.meanshape.view(1, -1, 3).mean(dim=1, keepdim=True)

q1: why ex_coeff use sum(1).mean() ,and other two use .sum()
`def reg_loss(id_coeff, ex_coeff, tex_coeff):

loss = torch.square(id_coeff).sum() + \
        torch.square(tex_coeff).sum() * 1.7e-3 + \
        torch.square(ex_coeff).sum(1).mean() * 0.8

return loss`

q2:why pose coeff(rot and tans) not include in reg_loss ?

q3:why illumination coeff（gamma）not include in reg_loss ?

Now I have a project on 3DMM re. I want to us BFM2019FaceModel and make my own facemodel_info.
May I get some way? THX

I used the bfm model with 53215 vertices and rewritten the point_ buf and skinmask in facemodel_info.mat run directly after that, but the results are not good. Here are some results. The texture seems to be fitted, but the shape and posture are very bad.

What can I do to rebuild a face with ears and neck? I would be very grateful if you could help me.

Hi, ascust, this repository is great. I am wondering how to modify the code to fit rgb+depth data using the real camera parameters?

Hi, I wonder if the repo can be adapted to other face models, such as FLAME?

In the code set focal=1015, Could you explain how to calculate the fov's value ?
And how to use the real fov's value of the whole picture not the crop's picture ?

Do you know the data pcaBasis and pcaVariance in BFM2019?
Could you please tell me how this data be calculateed?
Thank you very much!!!

I am going to render the face by BFM09ReconModel, using the following code,
pred_dict = self.recon_model(D3D_coeff, render=True)
where D3D_coeff is a tensor whose size is (1,257).
However, pred_dict['rendered_img'] that I finally get is a tensor whose elements are all 0, with H traceback "RuntimeError: tensor.H is only supported on matrices (2-D tensors). Got 4-D tensor."
I wonder what mistake I might make. I'd appreciate it if you could give me some advice...

I can't get the BFM 2009 anywhere. Have you consider to update your project?

Hello, thank you for this great work

I was wondering if there is anyway to make this code take an image folder as an input instead of one image at a time

I'm trying to use the fit_single_img.py but getting the above error. How can I solve this issue?

Traceback (most recent call last):
File "C:\Users\user\Downloads\3DMM-Fitting-Pytorch-master\3DMM-Fitting-Pytorch-master\3DMM-Fitting-Pytorch\convert_bfm09_data.py", line 105, in <m transferBFM09()
File "C:\Users\user\Downloads\3DMM-Fitting-Pytorch-master\3DMM-Fitting-Pytorch-master\3DMM-Fitting-Pytorch\convert_bfm09_data.py", line 32, in transferBFM09
original_BFM = loadmat('BFM/01_MorphableModel.mat')#loadmat
File "C:\Users\user\AppData\Local\Programs\Python\Python39\lib\site-packages\scipy\io\matlab\mio.py", line 223, in loadmat
MR, _ = mat_reader_factory(f, **kwargs)
File "C:\Users\user\AppData\Local\Programs\Python\Python39\lib\site-packages\scipy\io\matlab\mio.py", line 72, in mat_reader_factory
mjv, mnv = get_matfile_version(byte_stream)
File "C:\Users\user\AppData\Local\Programs\Python\Python39\lib\site-packages\scipy\io\matlab\miobase.py", line 231, in get_matfile_version
raise ValueError('Unknown mat file type, version %s, %s' % ret)
ValueError: Unknown mat file type, version 0, 0
PS C:\Users\user\Downloads\3DMM-Fitting-Pytorch-master\3DMM-Fitting-Pytorch-master\3DMM-Fitting-Pytorch> python convert_bfm09_data.py
Traceback (most recent call last):
File "C:\Users\user\Downloads\3DMM-Fitting-Pytorch-master\3DMM-Fitting-Pytorch-master\3DMM-Fitting-Pytorch\convert_bfm09_data.py", line 105, in
transferBFM09()
File "C:\Users\user\Downloads\3DMM-Fitting-Pytorch-master\3DMM-Fitting-Pytorch-master\3DMM-Fitting-Pytorch\convert_bfm09_data.py", line 32, in transferBFM09
original_BFM = loadmat('01_MorphableModel.mat')#loadmat
File "C:\Users\user\AppData\Local\Programs\Python\Python39\lib\site-packages\scipy\io\matlab\mio.py", line 223, in loadmat
MR, _ = mat_reader_factory(f, **kwargs)
File "C:\Users\user\AppData\Local\Programs\Python\Python39\lib\site-packages\scipy\io\matlab\mio.py", line 72, in mat_reader_factory
mjv, mnv = get_matfile_version(byte_stream)
File "C:\Users\user\AppData\Local\Programs\Python\Python39\lib\site-packages\scipy\io\matlab\miobase.py", line 231, in get_matfile_version
raise ValueError('Unknown mat file type, version %s, %s' % ret)
ValueError: Unknown mat file type, version 0, 0

Hello there!

First of all, I want to express my gratitude for your excellent work.

I have a question regarding the use of UV coordinates for BFM 09 models.

I have found a resource for UVs at https://github.com/anilbas/3DMMasSTN/blob/master/util/BFM_UV.mat, but I have encountered an issue.

The vertex numbers in the BFM 09 model and the UV resource don't seem to match.

I was wondering if you or anyone else has any suggestions on how to resolve this issue?