Johan Mathe
Johan Mathe
@shumatech Would really appreciate this PR - flashing on mac os X for a few architectures has been a huge pain.
Ok this is what I thought. I might play around with multi-gpu or TPU training with XLA to see if I can crank the batch size up. I am also...
Thanks for the insights!
Can you point me to the arguments you used and the file you used to produce this issue?
Thanks for the bug report! I will try to get to it.
Hey guys sorry this is pretty old, but I'm checking on the status for this, are you still interested in this PR to be merged?
Hard to parse from the logs here - it probably needs a few hours of investigation to see if it's possible to reproduce locally. I'm currently off for a week...