rhdong
rhdong
> Hi @rhdong , I fix some bugs(shape of TrainableWrapper) and build tf 2.4.0, based on your code. It seems the dynamic_embedding didn't updated in training process. > > Code...
> FYI, @kttian wrote a prototype for a differentiable hash map, roughly the equivalent of TensorList, as part of her internship project. Here's a colab that demonstrates direct gradient updates:...
Maybe ,you should remove the "tensorflow/stream_executor:stream_executor" and add "--config=cuda" to your bazeel build opt.: bazel build -c opt **--config=cuda** tensorflow/tools/pip_package:build_pip_package
To my best knowledge, TensorFlow has a warm start mechanism as a similar function. I think they have a meaningful design; maybe we can be inspired by it: https://github.com/tensorflow/tensorflow/blob/master/tensorflow/python/training/warm_starting_util.py#L419 I...
Hi @sivukhin, thank you for the feedback! We will give a resolution after the discussion. Thank you!
> @MoFHeka, thanks for quick reply! I will try to use Horovod, sure (I'm not familiar with it but looks like this is mature library which everyone used :-) )...
Hi @LinGeLin, it would be better to add a `[FEA]` to the comment.
Good job! Thank you @LinGeLin!
Hey @cjnolet @benfred, I implemented an initial version of `faster_dot` (without deep optimization), which could help improve performance in many classic scenarios. I will continue to communicate with cuSparse Team....