Faster gradient computation

polmorenoc / opendr

Modern OpenDR

141 stars 32 forks source link

Faster gradient computation #3

Open nicolasugrinovic opened 5 years ago

nicolasugrinovic commented 5 years ago

Thank you for releasing this! OpenDR uses chumpy for gradient computations which (if I am not wrong) run on CPU, is there a way to accelerate this? Have you tried to move this to Tensorflow, PyTorch or some other framework?

ahnHeejune commented 5 years ago

Yes, I thought the same thing as above. Do you think we can use tensorflow instead of chumpy? Both providing similar functions