YujiOshima commented on a change in pull request #9994: Tensorboard: Add
histogram callback
URL: https://github.com/apache/incubator-mxnet/pull/9994#discussion_r172404718
##########
File path: python/mxnet/contrib/tensorboard.py
##########
@@ -71,3 +72,43 @@ def __call__(self, param):
if self.prefix is not None:
name = '%s-%s' % (self.prefix, name)
self.summary_writer.add_scalar(name, value)
+
+ def node_histogram_visualization(self, prefix=None, node_names=None,
bins="auto"):
Review comment:
Thank you for comment.
This method can be used like below.
```python
lm = mx.contrib.tensorboard.LogMetricsCallback("/log")
model.fit(train,
...
epoch_end_callback=lm.node_histogram_visualization(),
...
)
```
I understand it would be nice that histogram callback is defined as a new
class.
But if you use both of scalar and histogram callback, you should use the
same `tensorboard.SummaryWriter` instance.
Another way, make a new class and give SummaryWriter instance as an argument
like below.
```python
class HistogramCallback(object):
def __init__(self, logging_dir, prefix=None, summarywriter=None):
```
Any ideas?
----------------------------------------------------------------
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
