Closed achalshah20 closed 3 years ago
If you could do that, I think it would be great!
On Fri, Jun 30, 2017 at 10:42 AM, Achal Shah notifications@github.com wrote:
I am using latest keras and tensorflow. I am getting many errors and many deprecation warnings. I can send PR with latest API changes for keras and tensorflow if you are interested!!!
— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/kuleshov/audio-super-res/issues/3, or mute the thread https://github.com/notifications/unsubscribe-auth/ACP_zgTljTZv-idXVFxZGvGf2tRqDyddks5sJTOLgaJpZM4OK3XS .
I would be interested in these changes as well. If you're not ready to merge the changes yet, I'd be happy to test them out locally first.
I did these changes locally, and should be able to push them soon after a few internal verifications.
That's great! I will try the changes out whenever they are ready.
Any news?
Any news? I am wondering these changes also!
Also very interested in this (as well as python 3 support). Any progress?
I'm currently trying to
No, I tried these codes in keras 2 and tensorflow 1.6, and I got a lot of errors which may be caused by the different version. Right now I am still working on it. I sent mail to the author but hasn't received a reply yet.
2018/04/10 午後0:01、Noah Bilgrien notifications@github.comのメール:
Also very interested in this (as well as python 3 support). Any progress?
— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/kuleshov/audio-super-res/issues/3#issuecomment-379958674, or mute the thread https://github.com/notifications/unsubscribe-auth/Afm4ewSUs3g_qVW8Wiz2ffiNZ7ossQU4ks5tnCCZgaJpZM4OK3XS.
I once tried to attempt this when tensorflow was around 1.2. But I couldn't able to reproduce the results so I am also waiting for official release from authors.
Me too. But it seems that the authors is too busy to push new program to github, so we have to wait.
在 2018年4月12日,上午9:52,Achal Shah notifications@github.com 写道:
I once tried to attempt this when tensorflow was around 1.2. But I couldn't able to reproduce the results so I am also waiting for official release from authors.
— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/kuleshov/audio-super-res/issues/3#issuecomment-380640203, or mute the thread https://github.com/notifications/unsubscribe-auth/Afm4e1coQws3vGOribGvGp4OimWuMis0ks5tnqVRgaJpZM4OK3XS.
I'm working on a version using Tensorflow 1.5, python 3, and zero Keras utilization. I'll publish and share it when I get it working.
Just on analyzing the code, I can't see how it ever worked. It doesn't seem to have ever increased the resolution of the output sounds beyond the input resolution.
I ‘ll appreciate it if you do that.
在 2018年4月12日,上午10:14,Noah Bilgrien notifications@github.com 写道:
I'm working a version using Tensorflow 1.5, python 3, and zero Keras utilization. I'll publish and share it when I get it working.
— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/kuleshov/audio-super-res/issues/3#issuecomment-380643395, or mute the thread https://github.com/notifications/unsubscribe-auth/Afm4e9VeDIWjcdQgluccqeKQkgzTssk0ks5tnqpXgaJpZM4OK3XS.
Hey polariss, I'm very close to having this working. Will likely publish something within the next few days, definitely within the next week. It doesn't have all the bells and whistles as this repo, but its at least functional.
That’s great! I’m looking forward to seeing your work. Please let me know when you publish it. Thanks
2018/04/16 午後2:27、Noah Bilgrien notifications@github.comのメール:
Hey polariss, I'm very close to having this working. Will likely publish something within the next few days, definitely within the next week. It doesn't have all the bells and whistles as this repo, but its at least functional.
— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/kuleshov/audio-super-res/issues/3#issuecomment-381483534, or mute the thread https://github.com/notifications/unsubscribe-auth/Afm4e-RFFFoGiW0SJWfgQzL0LFsfDsUYks5tpCulgaJpZM4OK3XS.
Hi kuleshov and Nobillygreen, I am a PhD student from the department of civil engineering and willing to apply deep learning to our side. Totally a newbie of Python and machine learning, I have learned some basics of python 3 and Keras. It is so exciting to find your paper Kuleshov, because it is strongly correlated to my research. Trying to understood and shift your code to Python 3 friendly but without any python 2 basis, I can only run the 'run.py' to the last line with an error: AttributeError: 'Namespace' object has no attribute 'func'. I am here seeking for some help, and also, is there any new progress with the new version? Many thanks to all of your work and kindly help. @kuleshov @nobillygreen
@nobillygreen did you ever make some progress you could share? Just wondering before I tackle the code myself..
@nobillygreen Also chiming in to ask if you had any luck with the new version, thanks!
Hey polariss, I'm very close to having this working. Will likely publish something within the next few days, definitely within the next week. It doesn't have all the bells and whistles as this repo, but its at least functional.
anyone made a updated version if this? @kuleshov did you made it? @nobillygreen ? did you manage it? @bafonso you tackled it yourself ? :D would be nice if someone can provice an updated version of this
i never did manage to get this working. Unfortunately I moved onto a different project. At this point, I'd recommend doing it in TF 2.X. Not sure I'll ever get around to it.
Me too, I have never run this project, can anyone provide code that can reproduce the paper?
Apologies for the very delayed response. I have updated the repo to run with TF 2.4. (Albeit with substantial use of tf.compat.v1
. . . fully refactoring the code is a larger project than I have time for at the moment.) I've verified that the example commands in the README run. Let me know if you run into any problems.
I am using latest keras and tensorflow. I am getting many errors and many deprecation warnings. I can send PR with latest API changes for keras and tensorflow if you are interested!!!