Is this project alive? about golgi

[chewxy]

Yeah it is... I made a mistake sometime in November - I coded most of Golgi in the vendor subpackage, and then proceeded to delete it. So it's taking some effort to get it back up. I'll pull up a copy tonight *Xuanyi Chew* *@chewxy on Twitter* *Read:* my blog <http://blog.chewxy.com> *[phone]*

…

On Mon, Jan 29, 2018 at 8:21 AM, Vlado Magyar ***@***.***> wrote: Has anyone started to implement this? It would be great to have a keras alternative to on gorgonia. — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub <#1>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AAczUlptB1GMt0WscaakrE-AqUfC3tdAks5tPOTmgaJpZM4Rv3Di> .

[chewxy]

@steele232 that's the plan. I wrote the stubs because I had expected me to finish porting over my personal library to this over the weekend. Two things prevented this:

1. Life got in the way (furniture building takes more hours than expected)
2. I added a new term to the calculus that defines the language here but when I put it thru Coq I couldn't prove correctness so I had to go back to the drawing board for a bit.

Having said that, the Coq bits don't really affect much. I'll try to get the other layers in.

Oops...
I also am interested.
The docs are very well written btw in the repos - hats off

[RameshRavone]

+1

[zhenghaoz]

+1

[steele232]

I'm excited to see this and try this out. I see some code committed 5 days ago but I haven't looked at it quite yet. Just the filenames look exciting.

[steele232]

I see that some of them are just stubs, but what I see looks good. I'm hoping this will be something that is super user friendly on top of gorgonia, kind of like Keras can be for Tensorflow.

[chewxy]

Closing this as this library is alive and has been used in prod. Though it is quite in an incomplete state (hence the lack of a tagged release). Will get tagged once LSTM and Conv are done