Is fine tuning on the pre-trained model provided a viable solution?

j96w / DenseFusion

"DenseFusion: 6D Object Pose Estimation by Iterative Dense Fusion" code repository

https://sites.google.com/view/densefusion

MIT License

1.09k stars 300 forks source link

Is fine tuning on the pre-trained model provided a viable solution? #133

Closed TommasoBendinelli closed 4 years ago

TommasoBendinelli commented 4 years ago

Hello, I would like to do pose estimation of custom workpieces. I was thinking, as we do for mask-cnn, if I could just download the pretrained model and then fine tune with my dataset, instead of training from scratch. Is it a viable approach?

j96w commented 4 years ago

Hi, yes, for different segmentation results finetuning, it is viable.

akeaveny commented 3 years ago

Hi @j96w, just to confirm there isn't a similar method to finetune results on a target domain with pose estimation?