skeleton-contrast
skeleton-contrast copied to clipboard
Comparison with supervised only training for AGCN
Hello,
Thank you for sharing this excellent work. I was wondering did you have the chance to do the "Comparison with supervised only training" in Table 7 for graph-based query encoder, i.e., f^STG(supervised-only) vs. f^STG(with-pretraining)? What the result could seem like?