search

Vegetebird / StridedTransformer-Pose3D

[TMM 2022] Exploiting Temporal Contexts with Strided Transformer for 3D Human Pose Estimation
MIT License
335 stars 37 forks source link

After training for several epochs, add refine module? #20

Closed haofeng0705 closed 1 year ago

haofeng0705 commented 1 year ago

is it refer to Exploiting Spatial-temporal Relationships for 3D Pose Estimation via Graph Convolutional Networks, the refinement module? how many epochs do I need to train to add this module？

Vegetebird commented 1 year ago

Our repo is built on top of ST-GCN, we train the first stage for 20 epochs.

haofeng0705 commented 1 year ago

thank you! btw, how can I save the predicted 3d poses and corresponding labels output from the model?

Vegetebird commented 1 year ago

You can save the outputs from the model (post_out) in https://github.com/Vegetebird/StridedTransformer-Pose3D/blob/6773a98f873f0d7bc6d8d0f6c0b1a478409e535e/demo/vis.py#L222