PortraitNet

PortraitNet copied to clipboard

Model training

3

Hello How are you? Thanks for contributing this project. I am training a new model with config model_mobilenetv2_with_two_auxiliary_losses and AISegment dataset for 8 days. But it is strange that the...

rose-jinyang

您提供的EG1800数据集labels怎么无标签？

5

您好，根据您提供的源码，本人已经全部调试完成，可以进行训练。但是为什么EG1800数据集的labels是全黑，而不是人像的掩膜呢？所以在测试时也只是显示背景，而没有人像显示？

carfei

Wrong format of pre-trained models

7

I can't unarchive them by tar or 7zip. Here are the log: 7z.exe a mobilenetv2_total_with_prior_channel.tar 7-Zip 19.00 (x64) : Copyright (c) 1999-2018 Igor Pavlov : 2019-02-21 Open archive: mobilenetv2_total_with_prior_channel.tar ERROR:...

SourceKim

when processing the images in supervisely_face_train_easy.txt and supervisely_face_test_easy.txt, I found these following images appear both in the train set and the test set. ``` [('Supervisely_face/SuperviselyPerson_ds4/img/pexels-photo-785667.png', 'Supervisely_face/SuperviselyPerson_ds2/img/pexels-photo-785667.png'), ('Supervisely_face/SuperviselyPerson_ds5/img/pexels-photo-356170.png', 'Supervisely_face/SuperviselyPerson_ds1/img/pexels-photo-356170.png'), ('Supervisely_face/SuperviselyPerson_ds5/img/pexels-photo-681294.png', 'Supervisely_face/SuperviselyPerson_ds1/img/pexels-photo-681294.png'),...

jonny4929

The mIoU is low

10

Thanks for your sharing @dong-x16 . I have an issue. When I train my model without the prior channel, the mIoU of the model can reach 94%, but when I...

jackieyung

resume

1

Thanks for contributing this project. When I was training, I changed the value of 'resume' to continue training, but an error occurred. I don’t know if there are any details...

Danelljian

wonder if you can provide the human dataset used in your third model

1

mobilenetv2_total_with_prior_channel(Training on Human with prior channel) which human dataset？ 谢了

tbcxx

HI,参照myTest目录下的video测试demo脚本, 下载页面提供的模型链接文件 将 154行 ：File = os.path.join(exp_args.model_root, 'model_best.pth.tar') 改成 bestModelFile = os.path.join(exp_args.model_root, 'mobilenetv2_total_with_prior_channel.tar') 运行后产生如下错误 videotest.py:101: YAMLLoadWarning: calling yaml.load() without Loader=... is deprecated, as the default Loader is unsafe. Please read...

chithize

模型测试为什么不能出结果？

1

感谢作者！本人利用您训练的模型pretrained_mobilenetv2_base.pth，进行图片和视频测试时，都无法成功测试，只显示了黑色背景，并没有将人像抠出来的实际效果图，然后我又看了您的EG1800数据集，labels标签都是黑色背景，并不是人像掩膜。希望作者看到后提供我一个合理的解释。谢谢！

carfei

Hi Thanks for your effort. Is there available google drive link for Supervisely portrait dataset? I cannot access in baidu drive...

junhwanjang