Closed ma3252788 closed 3 years ago
I am going to convert to PyTorch since it is the most dominant framework. It is a bit meaningless to release the code in a framework that few people are using, although I like MXNet very much.
But not in the near future. I am in quite a chaos and uncertain status recently.
If everything is set, I am planning to re-implement it in PyTorch in the coming spring festival.
I am going to convert to PyTorch since it is the most dominant framework. It is a bit meaningless to release the code in a framework that few people are using, although I like MXNet very much.
But not in the near future. I am in quite a chaos and uncertain status recently.
If everything is set, I am planning to re-implement it in PyTorch in the coming spring festival.
Okay, thank you very much, I look forward to seeing the pytorch version of aff in the near future~
Waiting for your pytorch version
@YimianDai Waiting for your pytorch version
@1187697147 Hi, @bobo0810 has contributed a Pytorch version https://github.com/YimianDai/open-aff/tree/master/aff_pytorch. Although it is not written by me, I hope it will help you.
As to me, currently, I am busy with finalizing my PhD Oral defense. I think I will have a lot of free time next week (I hope so) and will start to do all the things that I have ever promised in my replied issues of this repo.
Sorry for the delay.
@YimianDai Nice, good luck to you and thanks for your work!
Description:
Good luck!:smiley:
@bobo0810 @YimianDai Thanks for your great work, do you have pretrained models for pytorch version?
@bobo0810 @YimianDai Thanks for your great work, do you have pretrained models for pytorch version?
Unfortunately:pensive:, I may not be able to provide it. Because I only apply it to the company's private data set.
@hiyyg Hi, many thanks for being interested in my work. I am considering to fully convert to PyTorch and extend the conference paper to a journal article with more experiments.
If the accuracy of my new PyTorch implementation (not done yet) does not drop, I will release the pretrained models for PyTorch version. Please stay patient. I will update the process in this issue in the following days.
@hiyyg Hi, many thanks for being interested in my work. I am considering to fully convert to PyTorch and extend the conference paper to a journal article with more experiments.
If the accuracy of my new PyTorch implementation (not done yet) does not drop, I will release the pretrained models for PyTorch version. Please stay patient. I will update the process in this issue in the following days.
What's the result, please?
I am working on this. The next version will be PyTorch and Open-MMLab based.
I am working on this. The next version will be PyTorch and Open-MMLab based.
Got it. Thank you for your reply. Looking forward to your work
@hiyyg Hi, many thanks for being interested in my work. I am considering to fully convert to PyTorch and extend the conference paper to a journal article with more experiments.
If the accuracy of my new PyTorch implementation (not done yet) does not drop, I will release the pretrained models for PyTorch version. Please stay patient. I will update the process in this issue in the following days.
Hello, thanks for your good repository, Can you release pre-trained models for Pytorch?
请问有tensorflow版本的代码吗?
Or will you consider publishing the pytorch version of the code in the future?