search

see-- / keras-centernet

A Keras implementation of CenterNet with pre-trained model (unofficial)
MIT License
337 stars 84 forks source link

hello,Thank you for your code.I want to know when the training code will be updated. #3

Closed LCWdmlearning closed 5 years ago

LCWdmlearning commented 5 years ago

hello,Thank you for your code.I want to know when the training code will be updated. thanks a lot！！！

see-- commented 5 years ago

Probably within this month. It is not easy to build a fast training pipeline with keras. I.e. fit_generator & keras.utils.Sequence is slow. If you need a good (TensorFlow) object detector right now, take a look at tensorpack FasterRCNN: https://github.com/tensorpack/tensorpack/tree/master/examples/FasterRCNN. I will close this issue as it is tracked by #2.

LCWdmlearning commented 5 years ago

Thank you for your reply. Maybe it doesn't matter if it's slow. I'm going to retrain on a small training set. My own training code(keras) is not good， So I want to see your training code. Finally, thank you for your sharing.

------------------ 原始邮件 ------------------ 发件人: "see--"notifications@github.com; 发送时间: 2019年5月16日(星期四) 下午2:51 收件人: "see--/keras-centernet"keras-centernet@noreply.github.com; 抄送: "李灿伟"413645892@qq.com; "Author"author@noreply.github.com; 主题: Re: [see--/keras-centernet] hello,Thank you for your code.I want toknow when the training code will be updated. (#3)

Probably within this month. It is not easy to build a fast training pipeline with keras. I.e. fit_generator & keras.utils.Sequence is slow. If you need a good (TensorFlow) object detector right now, take a look at tensorpack FasterRCNN: https://github.com/tensorpack/tensorpack/tree/master/examples/FasterRCNN. I will close this issue as it is tracked by #2.

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub, or mute the thread.

white2018 commented 5 years ago

Thank you for your reply. Maybe it doesn't matter if it's slow. I'm going to retrain on a small training set. My own training code(keras) is not good， So I want to see your training code. Finally, thank you for your sharing. … ------------------ 原始邮件 ------------------ 发件人: "see--"notifications@github.com; 发送时间: 2019年5月16日(星期四) 下午2:51 收件人: "see--/keras-centernet"keras-centernet@noreply.github.com; 抄送: "李灿伟"413645892@qq.com; "Author"author@noreply.github.com; 主题: Re: [see--/keras-centernet] hello,Thank you for your code.I want toknow when the training code will be updated. (#3) Probably within this month. It is not easy to build a fast training pipeline with keras. I.e. fit_generator & keras.utils.Sequence is slow. If you need a good (TensorFlow) object detector right now, take a look at tensorpack FasterRCNN: https://github.com/tensorpack/tensorpack/tree/master/examples/FasterRCNN. I will close this issue as it is tracked by #2. — You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub, or mute the thread.

Thank you for your reply. Maybe it doesn't matter if it's slow. I'm going to retrain on a small training set. My own training code(keras) is not good， So I want to see your training code. Finally, thank you for your sharing. …

@LCWdmlearning Could you share your training code please? That would be appreciated.

Thanks a lot