You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Thanks for sharing these great projects.
I was wondering whether you could at some point update the available libraries for OSX?
I see them for r0.7 and r1.0. I have TensorFlow r.1.13 installed and somehow downgrading sounds more difficult than you’d think; pip not supporting versions of TensorFlow older than 1.13 anymore…
The text was updated successfully, but these errors were encountered:
djebel-amila
changed the title
TensorFlow r1.1/r1.3/r2.0
libraries for OSX × TensorFlow r1.1/r1.3/r2.0
May 6, 2019
Hi, yes sorry for being so behind on this, I've been busy with other things :/
I believe there is no GPU support on OSX anymore, so I've stopped supporting OSX I'm afraid (also, I don't have an OSX computer anymore, so I couldn't do it even if I wanted to).
Though I should say, technically it's not essential that the ofxMSATensorflow tensorflow library version is the same as your pip version, since they are stored completely separate places. However of course, it is almost certain that if you build and export a tf model in python with tf1.13, then you will not be able to load it with ofxMSATensorflow tf 0.7 since there will be a huge mismatch in compatibility between 0.7 and 1.13. However if you're using just some common basic operators, then you may be able to load your 1.13 model with 1.0 (though I can't garantee that of course).
No worries, and thanks. Another push (I needed) to go for Linux.
Just in case, gilbertfrançois has a library for tensorflow 1.11 for linux/osx: https://github.com/gilbertfrancois/ofxGFTensorflow/releases
Would it not fit ofxMSATensorflow as well?
Hi Memo,
Thanks for sharing these great projects.
I was wondering whether you could at some point update the available libraries for OSX?
I see them for r0.7 and r1.0. I have TensorFlow r.1.13 installed and somehow downgrading sounds more difficult than you’d think; pip not supporting versions of TensorFlow older than 1.13 anymore…
The text was updated successfully, but these errors were encountered: