search

researchmm / PEN-Net-for-Inpainting

[CVPR'2019] PEN-Net: Learning Pyramid-Context Encoder Network for High-Quality Image Inpainting
https://arxiv.org/abs/1904.07475
MIT License
361 stars 77 forks source link

pre-trained model for places2 #9

Open ds-za opened 4 years ago

ds-za commented 4 years ago

Hi - Very nice piece of work. While I can access the pre-trained models CELEBA-HD, DTD and Facade, it does not work for Places2. Might be related to access rights for the Google Drive. Could you please check? Thanks

sadbhawnathakur commented 4 years ago

Hi - The same issue is with me while trying to download pre-trained model for Places2.

zengyh1900 commented 4 years ago

Thank you for your reminder! I have just noticed it. I will upload it asap after checking. Thanks!

ofirkris commented 4 years ago

@1900zyh Thanks - it's the best model to test with

zengyh1900 commented 4 years ago

Hi all, thank you for your interest. I have updated the driver, you can check it now.

zhengbowei commented 4 years ago

Hi,everybody! I want to test the pretrained model, but I do not know how to prepare "zip_root": "../datazip" and the setting of "-l", thank you for your answer!

m2005dh commented 4 years ago

@1900zyh Hello, Dear zyh, I also have a problem to set the parameters flist_root and zip_root. I will be appreciated to anyone help me with this issue especially @1900zyh. Thank you

micklexqg commented 4 years ago

the same problem when prepare "zip_root": "../datazip" , so how to get or create them? @1900zyh

ZHANG-hengzhi commented 3 years ago

大家好！ 我想测试预训练的模型，但我不知道如何准备“zip_root”：“./datazip”和“-l”的设置，谢谢您的回答！

你好，请问你现在会配置自己的数据集了吗，我也是zip_root不知道是什么样子的