About train.py - parameterization
ryukyungjoon opened this issue · 1 comments
Dear Vivek,
Thank you for the help with my previous questions regarding train.py. I have one more question about it.
Q. In a scenario where multi-processing is not possible due to hardware limitations and training needs to be done with only 1 GPU, how can the training be performed?
parameterizations = [
"se3_log_map",
"so3_log_map",
"axis_angle",
"euler_angles",
"quaternion",
"rotation_6d",
"rotation_10d",
"quaternion_adjugate",
]
Should I pass these parameters to the main() function with a for loop and train them sequentially? The essence of the question is how to overcome the situation when hardware specifications do not allow multi-processing.
And, What is mean each parameterization item.
Thank you in advance for your response.
hi @ryukyungjoon , you don't need one model for each parameterization. in diffpose, we just train using se3_log_map
. the parameterizations are tested when running test-time optimization. the multiprocessing here is used to train a separate model for each subject in parallel on multiple GPUs.