vrv

6 ms·

31.
by vrv 11y ago
Thank you for your work on putting Keras on top! To answer your questions: - We don't (yet) have a tensor contraction op -- just a matter of getting some dev time to call the existing Eigen contraction code in an op. Hopefully in the
32.
by vrv 11y ago
FYI, we're building the pip packages today -- we'll send out an announcement to the TensorFlow discussion list and update the website when things are actually done. :)
33.
by vrv 11y ago
Thanks, would be good to have multiple sources of verification that HEAD now supports this natively without issues like unexpected NaNs. https://github.com/tensorflow/tensorflow/issues/25#issuecomm... is one
34.
by vrv 11y ago
We're looking to support Python 3 -- there are a few changes we are aware of that are required, and we welcome contributions to help! Tracking here: https://github.com/tensorflow/tensorflow/issues/1
35.
by vrv 14y ago
The degenerate/extreme/unrealistic case of Incast is you have switch buffer capacity to store N segments, you talk to M servers that each return one segment, and M >> N. Although RTO is calculated based on RTT and RTTVAR, in the ex
36.
by vrv 15y ago
I'll let others chime in, but the DCTCP paper mentions "our application reduces the amount of data each worker sends and employs jitter. Facebook, reportedly, has gone to the extent of developing their own UDP-based congestion control [29].
37.
by vrv 15y ago
This work is really cool and is attacking an important problem in a lot of datacenter networks. It's also great that they've released the source code too. As someone who's worked on this problem, I encourage those interested in reading up