SoftRas icon indicating copy to clipboard operation
SoftRas copied to clipboard

Project page of paper "Soft Rasterizer: A Differentiable Renderer for Image-based 3D Reasoning"

Results 56 SoftRas issues
Sort by recently updated
recently updated
newest added

Can you share the code for the pose optimization example with the cube? Thanks!

After I finished processing dependencies for soft-renderer==1.0.0, I run you code. But I has a error below. I try python=3.6 or python=3.7. it all failed. my cudatoolkit =10.1.243 . Traceback...

When running the function: `faces_info, aggrs_info, soft_colors = \ soft_rasterize_cuda.forward_soft_rasterize(face_vertices, textures, faces_info, aggrs_info, soft_colors, image_size, near, far, eps, sigma_val, ctx.func_dist_type, ctx.dist_eps, gamma_val, ctx.func_rgb_type, ctx.func_alpha_type, ctx.texture_type, fill_back)` Got the error message:...

There are serveral images in the attachment that illustrate the program execution progress and the errors. Firstly, I create the virtual environment with anaconda, 4.12.0 version . And I follow...

my code environment is: ubuntu+pytorch1.1.0+cuda10.0. I can install and run 0th(rendering) successfully, but when I run this: CUDA_VISIBLE_DEVICES=0 python examples/recon/train.py -eid recon, I meet this problem. Anybody else meets the...

I have successfully built Detectron2 under this environment. But when i build SoftRas,i meet this problem. Can you give me some advice? Thanks so much. 1 error detected in the...

How to use opencv style matrix instead of OpenGL style? could you provide a demo?

Hello, when I use 'python setup.py install', I got this problem: `g++: error: SoftRas-master/build/temp.linux-x86_64-3.8/soft_renderer/cuda/load_textures_cuda.o: No such file or directory g++: error: SoftRas-master/build/temp.linux-x86_64-3.8/soft_renderer/cuda/load_textures_cuda_kernel.o: No such file or directory error: command 'g++'...

In your paper, you describe Implement Detail for 642 vertices. What is the limit on the number of vertices this network can handle?

It would be good to add a libtorch C++ API to this library. For example, right now I need a differentiable renderer with C++ physics simulation code, and mixing Python...