search

Xinyu-Yi / TransPose

A real-time motion capture system that estimates poses and global translations using only 6 inertial measurement units
https://xinyu-yi.github.io/TransPose/
GNU General Public License v3.0
383 stars 75 forks source link

How do you process variable length input sequences while training? #9

Closed VimalMollyn closed 2 years ago

VimalMollyn commented 2 years ago

Hi! Congrats on the amazing paper and thanks so much for providing your source code! I was wondering if you guys used variable input sequences while training and if you could provide code for handling variable input sequences?

Thanks!

Xinyu-Yi commented 2 years ago

The training sequences are clipped into short ones in lengths <= 300 for training. You can discard the short sequence (< 300) for simplicity. Or you can use the functions in torch.nn.utils.rnn to pack the sequences of different lengths (which is our implementation).

VimalMollyn commented 2 years ago

Great, Thanks!