MINJUNG KIM

icon company @DAVIAN-Lab icon location 3D Vision and Video Generation

Results 21 comments of MINJUNG KIM

about main table 1

I have not changed the script from the original one. Actually, the branch that I used is the default 'custom' branch. Could it be the reason? Or any version subtleties?

Testing Trained Weights

I tried "Python 3.5, Keras 2.0.8, and Tensorflow 1.3.0" setting, but failed, yielding "tensorflow has no attribute math".

Result on MOT Benchmark site

never mind, I found it by changing filter option to 'all'

getting the following error when building nimble in conda enviornment FAILED: caffe2/CMakeFiles/torch_cuda.dir/utils/torch_cuda_generated_math_gpu.cu.o subprocess.CalledProcessError: Command '['cmake', '--build', '.', '--target', 'install', '--config', 'Release', '--', '-j', '8']' returned non-zero exit status 1.

I added the following to caffe2/utils/math_gpu.cu ``` #include ``` and it worked like charm. ref : https://github.com/pytorch/pytorch/issues/61011

Input shape issue and lack of bias.

I think iterating over timesteps seems reasonable

About Synthetic Defous Blur Data

Also, the image folder of defocuspool data is empty.

RuntimeError: No ffmpeg exe could be found. Install ffmpeg on your system, or set the IMAGEIO_FFMPEG_EXE environment variable.

I got a hint from @hmaarrfk and tried os.environ["OMP_NUM_THREADS"]=1 and this solved my problem. Thanks.

nvcc fatal : Unsupported gpu architecture 'compute_86'

By modifying the following, ``` TORCH_CUDA_ARCH_LIST='8.6+PTX' ``` I was able to avoid the 'compute_86' error message. But now I get new error messages. ``` /tmp/ccLFVASL.s: Assembler messages: /tmp/ccLFVASL.s: Fatal error:...

about noisy results from own dataset

same here

Ray distribution loss (sigma_loss) is not used?

I have the same question