Hubrt Hui
Hubrt Hui
i found the 'inference': 'oup_dim' in __config__(pose.py) is 68,however the loss is put on the first 34-d of the output of the network? It would be appreciate that you could...
dear author: did you train without segmentation?