search

HYOJINPARK / ExtPortraitSeg

62 stars 11 forks source link

hi, where is the trainning code? #1

Open zhangyunming opened 4 years ago

zhangyunming commented 4 years ago

good joob ,but i have not seen the code which you said in your 'readme'.

HYOJINPARK commented 4 years ago

Thanks for your comments. This work have sponsored by Naver Corporation. Today all of the security checks are ended and I submit code. A member from Naver will upload our code soon and I will notice the link as soon as possible. The final code will be released in Naver Clova AI open GitHub. I am sorry to delay and thanks for your patients.

HYOJINPARK commented 4 years ago

The code is released!! Thanks for your waiting. https://github.com/clovaai/ext_portrait_segmentation

zhangyunming commented 4 years ago

hi, @HYOJINPARK ,thank you ,i have seen the project. and is there a demo.py that read a pic or camera?

HYOJINPARK commented 4 years ago

Sorry for late answer You can check etc/Visualize_video.py

yasar-rehman commented 4 years ago

Hi @HYOJINPARK,

Thank you for sharing the code.

I have tried the etc/Visualize_video.py file, however, I couldn't find the model weights "../result/Dnc_SINet11-24_2218/model_3.pth". I have also tried with the ExtremeC3.pth and SINet.pth, but it's giving me a dimension mismatch between the model and the weights. Could you guide me in solving this issue?

HYOJINPARK commented 4 years ago

@yasar-rehman Thanks for your asking. The reason of mismatching dimension is with or without Lavasz loss If you used the SIN.pth file with ./setting/Test_SINet.json You can run it

The files are in here https://github.com/clovaai/ext_portrait_segmentation

yasar-rehman commented 4 years ago

@HYOJINPARK Thank you for the help. :)