CNevd commented on a change in pull request #7082: Sparse Tensor: request for
reviews
URL: https://github.com/apache/incubator-mxnet/pull/7082#discussion_r133667882
##########
File path: python/mxnet/model.py
##########
@@ -113,25 +127,36 @@ def _update_params_on_kvstore(param_arrays, grad_arrays,
kvstore, param_names):
# push gradient, priority is negative index
kvstore.push(name, grad_list, priority=-index)
# pull back the weights
- kvstore.pull(name, arg_list, priority=-index)
+ if _contains_non_default_storage(arg_list):
Review comment:
Consider arg_list has 1 row_sparse weight and many dense weights,
```_contains_non_default_storage(arg_list)``` will be true, so user have to
call ```kvstore.pull``` for all the dense weights manually out of
```model.py```. maybe there is some way to split them (row_sparse and dense)
so we can only pull row_sparse weights manually?
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org
With regards,
Apache Git Services
