TnoobT
TnoobT
use this version yolov8 https://pypi.org/project/ultralytics/8.0.132/#files, no thanks
@VainF 大佬求回复
@VainF 大佬，这是我测试剪枝的结果。https://blog.csdn.net/qq_33596242/article/details/133774348 ，您能帮我看看正常不。正常的话，我这流程也可以给其他人做个参考
@Reaidu 用我博客里的版本
@huangzongmou 请问直接剪枝和稀疏训练后再剪枝，精度差多少
> @TnoobT What was the batch size? Did you train the model? Full model? On only decoder layer? Thanks for the info. No changes made, consistent with the training code...
> @TnoobT can you please share the training results? Did you train on custom dataset? what about the loss? is it decreased? The effect is not good, the inpaint is...
hi, can you solve Converting Output SMPL to the standard SMPL parameters。 i used Rh replace pose[:3] and use Th replace trans, but results seem not correct.