Is there an overview, how the config object for the constructors of the neural network are defined?

This would be interesting:

• Has brain.NeuralNetwork and brain.recurrent.RNN the same config attributes?
• What is the description of the inputSize attribute?

Thank you for the great work with this library, it is realy amazing!

What do you want to do with Hls.js?
I want to use hlsjs with DRM (widevine, and maybe others if available), but I was not able to any documentation around how to do it.

What have you tried so far?
to set the following config properties:

• widevineLicenseUrl: this.drmProviders.widevine
• emeEnabled: true

But I need to modify the license request to inject our specific token.

Where can I find information on DISH, RADIO, GATHER and SCATHER draft socket types?

From my understanding the RADIO-DISH pattern is like the PUB-SUB with thread-safety, better performance and without SNDMORE. My guess would be that GATHER-SCATHER acts like XPUB-XSUB with the same tradeoffs?

Currently you can only download songs by setting them as the active song and using the top-right menu, or through the main UI section. Being able to download directly from the queue through the song's context menu would be nice

Once the v2 becomes to be stable, we need to update the README and the examples provided within the project.

I think the documentation must include a note on the diamond flow and how xstream manage it, something like an explanation of atomic updates of flyd.

As well described by @staltz in this answer would be best to avoid diamond shaped stream graph:

![ima

When dealing with a database with a large number of columns and tables, it would be handy to hide all disabled columns to make the tree more comprehensible. A suggestion would be to add a small toolbar above the tree with a toggle button to hide/view disabled columns. Alternatively it could be an option in the context menu.

Adding information on how to test Kefir Observables would be a nice addition to the docs. Wondering if we should put all of that into the main docs site, or have a separate site for testing with chai-kefir, or maybe moving all the docs into a kefirjs.github.io repo & website. Thoughts?

现在很多新的项目都是使用Swift，RxSwift学习曲线较大，在寻找替代方案，很期待你们的Swift支持！

有具体的时间计划吗？谢谢！

The docs say that "user-provided effects that are passed to eval are never interrupted once they are started" (here). As far as I can tell, this is not true.

Just solved an issue for a user that was confused by form-data's error message. In the following case:

const FormData = require('form-data');

const form = new FormData();
const user = { avatar_file: 'abc.png' }; // no nickname in here

form.append('nickname', user.nickname); // <- you're passing undefined here

The error thrown

I got this error - TypeError: (("document.querySelector("#react-view").setAttribute("data-react-checksum", "" + (intermediate value)) + "")") is not a function

How can I fix it?

It sounds as though there is desire from users to have a MP4 concat feature (to include myself). Specifically I'm talking about a mapping that describes multiple clips that should be concatenated into a single MP4.

It would also be nice to see further documentation on the current concat filter implementation. When I think of the concat filter I'm thinking of the ffmepg concat filter, or demuxer

Streams is a bit different in structure from other specifications in that it uses a heading for every API surface and abstract operation. This leads to a much larger table of contents than other specifications, and as such, I think the TOC is not very useful in its current form.

I think we should omit:

• Everything under "Class X"
• Every abstract operation (but keep the categories, e.g.

The architecture mentioned R API support, but I was not able to find documentation about it a all. Would you consider creating some sample recipes for R?

Thanks!