Where communities thrive


  • Join over 1.5M+ people
  • Join over 100K+ communities
  • Free without limits
  • Create your own community
People
Activity
  • 20:10
    aaronmondal commented #2264
  • 20:03
    aaronmondal commented #2263
  • 16:37
    aaronmondal synchronize #2265
  • 16:17
    aaronmondal commented #2252
  • 16:16
    aaronmondal commented #2252
  • 16:02
    aaronmondal commented #2265
  • 16:00
    aaronmondal synchronize #2265
  • 15:52
    aaronmondal closed #2256
  • 15:52
    aaronmondal commented #2256
  • 15:47
    bot-of-gabrieldemarmiesse commented #2265
  • 15:46
    google-cla[bot] labeled #2265
  • 15:46
    boring-cyborg[bot] labeled #2265
  • 15:46
    aaronmondal review_requested #2265
  • 15:46
    aaronmondal review_requested #2265
  • 15:46
    aaronmondal opened #2265
  • 14:53
    bot-of-gabrieldemarmiesse commented #2264
  • 14:52
    google-cla[bot] labeled #2264
  • 14:52
    boring-cyborg[bot] labeled #2264
  • 14:52
    aaronmondal opened #2264
  • 14:48
    bjtho08 edited #2262
Harshith MohanKumar
@Harsh188
Screenshot 2020-09-21 at 8.21.42 PM.png
bhack
@bhack
@Harsh188 Is the same issue you had at tensorflow/addons@51d8eb6
Harshith MohanKumar
@Harsh188
Yeah, I realized but I didn't see the connection between the two. It's pretty weird how a couple of tabs can fix the error.
divinereincarnation
@amishra123
hi, does anyone know how to look at assets under a model using REST API for tensorflow/serving just like /v1/models/xxx/metadata?
please and thank u
i believe it's not supported
Jason Zaman
@perfinion
heya, addons fails to build for my 3090, i need some updates to the cuda_configure stuff to work with cuda11.1 , wanted to ask a few things before I send the PR
I pulled in the new find_cuda_config.py from TF (and reformatted with black obvs), it pulls in the cusparse bits it needs, but it also adds a bunch of other changes, is that okay or should I only add the minimal needed?
it mostly just adds a few more search paths and stuff like that which i guess isnt strictly required but i figure keeping in sync with TF is good?
Jason Zaman
@perfinion
cuda_configure.bzl has diverged a whole ton from TF's one, a lot is to do with the remote build stuff so that one I added only the cublas bits i needed
bhack
@bhack
@perfinion We could try and how our CI will work with this change... Can you open a PR?
Jason Zaman
@perfinion
bhack: sounds good, i'll open it and we do review there then yeah
it seems to work well enough to me, so yep as long as y'all have no immediate objections i'll open the PR :)
bhack
@bhack
@perfinion If you want to check last switch PR
Jason Zaman
@perfinion
oh cool, that helps for reference yeah
Harshith MohanKumar
@Harsh188
@perfinion I've been having trouble getting my 3080 to work with TF. Are you using CUDA 11.1 with cuDNN v8?
Jason Zaman
@perfinion
yep, the RTX30xx requires CUDA11.1 and cudnn8, you cant use 11.0
ive got everything working to train my stuff, gonna fix up comments on the PR to get it merged but it does work currently for me
Philip May
@PhilipMay
Does the RTX30xx need 300 Watt or 250?
Jason Zaman
@perfinion
they're 350w, so neither :P
i wont be cold this winter for sure
bhack
@bhack
I think 3080 Is 320W
Jason Zaman
@perfinion
yeah it is
bhack
@bhack
3070 Is 220W
Only 3090 Is 350W
Jason Zaman
@perfinion
yup
also only the 3090 has nvlink, the 3080 and 3070 dont have the connectors
Harshith MohanKumar
@Harsh188

Hello! Does anyone here use docker to run TensorFlow by any chance?
I've been trying to mount a local folder that contains my Jupyter notebooks by running:

docker run -it --rm --name tf -v $(pwd):/notebooks -p 8888:8888 -p 6006:6006 tensorflow/tensorflow:latest-gpu-jupyter

but when I open up my localhost the only files which are available are the tensorflow-tutorials. I followed a youtube tutorial that ran the same command that I am currently running and it seemed to work just fine for him. I tried to go through and understand how mounting works to see if that's what my issue is but since the command doesn't result in an error, therefore, I'm confused why my files won't show up. Also If I wanted to use TensorFlow addons would I have to create another image? I apologize in advance since this isn't directly related to addons discussions.
I haven't defined a Docker file. I'm not sure if this is what I'm missing, but the person in the yt tutorial didn't so I assumed that It's not needed.

bhack
@bhack
We already have Tensorflow Addons images
And we also have Vscode integration for these images
(if you want to develop with TF Addons in docker)
Harshith MohanKumar
@Harsh188
@bhack I can't seem to find any Addons images on docker hub. Also, Addons isnt listed as a repo under the parent TensorFlow organization.
bhack
@bhack
Harshith MohanKumar
@Harsh188
Thanks! Also, is there a reason why tfaddons isn't listed under tensorflow in docker hub?
Sean Morgan
@seanpmorgan

The future is ARM :)

Happy to see TFA being compiled and included in the Apple fork:
https://github.com/apple/tensorflow_macos

Kiran Khanna
@kiran-glitch
Hi everyone , I am Kiran Khanna , am from India , I am interested in contributing in Tensorflow .
Sean Morgan
@seanpmorgan
Welcome Kiran! Check out our contributing guidelines and let us know if you have any questions or ways we can improve it https://github.com/tensorflow/addons/blob/master/CONTRIBUTING.md
Kiran Khanna
@kiran-glitch
Thank you , I will check it out
aaryaab
@aaryaab
Hello everyone! I am Aarya, I am from India. I have some experience in open source contribution. I look forward to make some valuable contribution in Tensorflow!
Vishesh Tripathi
@Vishesht27
Hi everyone , I have been contributing to open source and I want to contribute to Tensorflow. Can someone please tell me the technologies I need to learn for starting contributing to Tensorflow
bhack
@bhack
Dear Tensorflow or Tensorflow addons?
Vishesh Tripathi
@Vishesht27
I have basic knowledge of Python, C++, java and ML. Is that good for me start?
bhack
@bhack
You can start with TF tutorials
Vishesh Tripathi
@Vishesht27
Oh,on it