How to save as .obj

[jay-thakur]

Hi,

I used this code to save as .obj-

import chainer
import numpy as np
import scipy.misc
import neural_renderer

renderer = neural_renderer.Renderer()

vertices, faces, textures = neural_renderer.load_obj( '/home/user/jay/straight.obj', load_texture=True, texture_size=16)

neural_renderer.save_obj('/home/user/jay/chainer_save.obj', vertices, faces, textures)

But facing issue in renderer.py. So i changed below code like this.
faces = torch.cat((faces, faces[:, list(reversed(range(faces.shape[-1])))]), dim=1).detach()

Then i was getting error in vertices_to_faces.py so i changed below vertices & faces assert to 2.
assert (vertices.ndimension() == 2)
assert (faces.ndimension() == 2)

But I am getting error at
assert (vertices.shape[0] == faces.shape[0]) because
vertices.shape[0] = 34817
faces.shape[0] = 69630.

Where am i doing wrong ? Could you please help me to fix this?

[minghaoPA]

I think the shape[0] is to make sure batch sizes are equal to each other. So you need to add an extra '0' dimension as batch size, like line25 and line26 in example3.py.

    vertices, faces = neural_renderer.load_obj(filename_obj)
    self.vertices = vertices[None, :, :]
    self.faces = faces[None, :, :]

Hopefully my understanding is helpful to your problem.

And if you have time, please also take a look at my question #21 , which is also related to save_obj.py.

[jay-thakur]

I believe, if we add extra '0' as you suggested, then vertices & faces will be of 3 dimension. & in save_obj() we have assert statement for vertices & faces to 2. Please correct me if i am wrong.

I am expecting to save the .obj with its texture . Do you have working code for the same? Could you please share your example3.py ?

Thanks,
Jay

[minghaoPA]

Hi Jay. For your concern, please check out my code at question #21. I am sharing my piece of code there.

But anyway, I just slice out the first dimension and feed the rest into save_obj. And I am also a little confused about your question. The vertices_to_faces is not called either in load_obj.py or save_obj.py. So I don't encounter any problem related to dimension assertion. Maybe it's because you are using torch edition from another repo, not exactly the same code as this repo?

[minghaoPA]

For simplicity, here are the codes added before #draw object module in example3.py

    # save obj with textures
    #TODO: textures are not reasonable
    model.textures_1 = cf.concat((model.textures, model.textures.transpose((0, 1, 4, 3, 2, 5))), axis=1)
    neural_renderer.save_obj('result.obj', model.vertices[0], model.faces[0], cf.tanh(model.textures_1[0]).array)

[jay-thakur]

Sorry, I am using another repo.

[Goldname100]

@jay-thakur Which repo are you using?

[kelly08385]

I have the same problem when saved the rendered object.
The error said

cupy.cuda.compiler.CompileException: /tmp/tmp72lc77x3/44a6062960e6de87983184b4a18c1a10_2.cubin.cu(20): error: a value of type "const float *" cannot be used to initialize an entity of type "float *"

/tmp/tmp72lc77x3/44a6062960e6de87983184b4a18c1a10_2.cubin.cu(21): error: a value of type "const float *" cannot be used to initialize an entity of type "float *"

/tmp/tmp72lc77x3/44a6062960e6de87983184b4a18c1a10_2.cubin.cu(75): error: expression must be a modifiable lvalue

I think it's chain, cupy or cuda versions problem.
I run in python3.8 and cuda10.2
What chain and cupy version should I select?
thanks

[caojiehui]

I have the same problem when saved the rendered object. The error said

cupy.cuda.compiler.CompileException: /tmp/tmp72lc77x3/44a6062960e6de87983184b4a18c1a10_2.cubin.cu(20): error: a value of type "const float *" cannot be used to initialize an entity of type "float *"

/tmp/tmp72lc77x3/44a6062960e6de87983184b4a18c1a10_2.cubin.cu(21): error: a value of type "const float *" cannot be used to initialize an entity of type "float *"

/tmp/tmp72lc77x3/44a6062960e6de87983184b4a18c1a10_2.cubin.cu(75): error: expression must be a modifiable lvalue

I think it's chain, cupy or cuda versions problem. I run in python3.8 and cuda10.2 What chain and cupy version should I select? thanks

Hello, I have encountered the same issue as you. How did you resolve this problem? Was it related to the version?