Hello,
Considering your amazing efficiency on pandas, numpy, and more, it would seem to make sense for your module to work with even bigger data, such as Audio (for example .mp3 and .wav). This is something that would help a lot considering the nature audio (ie. where one of the lowest and most common sampling rates is still 44,100 samples/sec). For a use case, I would consider vaex.open('Hu

https://github.com/infinitered/nsfwjs/blob/ebcd41c46087a3f42c6577f96acc53d7a934b068/src/index.ts#L68

Hello, it seems, although not explicit I can save the model to different schemas by referencing the underlying "model" attribute in the model returned by nsfwjs.load() e.g.

`nsfwjs.load(path).then(function (newModel) {
console.log("path", path);
if(newModel) {

How to use Watcher / WatcherClient over tcp/ip network?

Watcher seems to ZMQ server, and WatcherClient is ZMQ Client, but there is no API/Interface to config server IP address.
Do I need to implement a class that inherits from WatcherClient?

https://igel.readthedocs.io/en/latest/_sources/readme.rst.txt includes a link to the assets/igel-help.gif, but that path is broken on readthedocs.

readme.rst is included as ../readme.rst in the sphinx build.
The gifs are in asses/igel-help.gif

The sphinx build needs to point to the asset directory, absolutely:

.. image:: /assets/igel-help.gif

I haven't made a patch, because I haven't

Hi, Thanks for the awesome library!

So I am running a Kmeans on lots of different datasets, which all have roughly four shapes, so I initialize with those shapes and it works well, except for just a few times. There are a few datasets that look different enough that I end up with empty clusters and the algorithm just hangs ("Resumed because of empty cluster" again and again).

I conceptually

Bounds check and call [] operator

Right now, these aren't caught until we try to gob-encode. Consider failing faster in type-checking to avoid too much confusion/loss when it works with local execution.