search

hjwdzh / PrimitiveNet

PrimitiveNet: Primitive Instance Segmentation with Local Primitive Embedding under Adversarial Metric (ICCV 2021)
MIT License
62 stars 11 forks source link

Wrong pre-trained model? #4

Closed MulinYu closed 2 years ago

MulinYu commented 2 years ago

Hello,

I tried using your provided ABC pre-trained model to test your ABC data. The results are bad, I think that you maybe provide the wrong pre-trained model.

It's normal that after unzipping the 'abc.pth' file, there is a ' scene.pth' file in the checkpoints folder?

Thanks in advance. Best. Mulin

MulinYu commented 2 years ago

The output of evaluation is as follow:

0.5 0.00517630176301763 0.55 0.0043819188191881915 0.6000000000000001 0.003664411644116441 0.6500000000000001 0.003254407544075441 0.7000000000000002 0.0029212792127921283 0.7500000000000002 0.0026137761377613778 0.8000000000000003 0.0024600246002460025 0.8500000000000003 0.0022037720377203773 0.9000000000000004 0.0021781467814678147 0.25 0.04440856908569086

hjwdzh commented 2 years ago

This is definitely incorrect. Did you find the pretrained model which is named abc-pretrained.pth? If not, probably I did not upload the correct one... Let me know. If not, I will re-upload the model.

hjwdzh commented 2 years ago

I just verified and confirmed that my uploaded one is broken... I will upload a new one in 3-5 days. Will let you know when it is ready.

MulinYu commented 2 years ago

I just verified and confirmed that my uploaded one is broken... I will upload a new one in 3-5 days. Will let you know when it is ready.

Thanks a lot.

hjwdzh commented 2 years ago

Please view the new link in download.sh.