You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@mxnet.apache.org by "Gautham Kollu (JIRA)" <ji...@apache.org> on 2018/09/10 23:54:00 UTC
[jira] [Created] (MXNET-938) RNN.hybrid_forward() spends a
significant time on _rnn_param_concat
Gautham Kollu created MXNET-938:
-----------------------------------
Summary: RNN.hybrid_forward() spends a significant time on _rnn_param_concat
Key: MXNET-938
URL: https://issues.apache.org/jira/browse/MXNET-938
Project: Apache MXNet
Issue Type: Improvement
Components: Gluon
Reporter: Gautham Kollu
Attachments: gluon_lstm.py
Issue :
RNN.hybrid_forward() spends a significant time on _rnn_param_concat
Root Cause:
Creating a simple LSTM (256 -> 256) with gluon (gluon_lstm.py) and running inference on CPU with the profile shows the following output. \{ Refer gluon_lstm.py for the code}
_rnn_param_concat() takes more time than the RNN operator.
| | | | | | | | | | |
|-------------------|-------|--------|-----------|----|------------|----|------------|----|------------|
|operator| | | | | | | | | |
|=================| | | | | | | | | |
|Name|Total|Count|Time (ms)|Min|Time (ms)|Max|Time (ms)|Avg|Time (ms)|
|----|----|------|---------|—|----------|—|----------|—|----------|
|RNN| |72|72.312| |0.065| |33.15| |1.0043|
|_zeros| |788|1.76| |0.001| |0.011| |0.0022|
|DeleteVariable| |1110|0.637| |0| |0.002| |0.0006|
|SwapAxis| |462|2.223| |0.003| |0.015| |0.0048|
|_rnn_param_concat| |76|103.813| |1.122| |2.08| |1.366|
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscribe@mxnet.apache.org
For additional commands, e-mail: issues-help@mxnet.apache.org