A GUI that can display Leela's chances of winning in every single position

Imagine, your chances of winning can be visually seen on the board?

This GUI can now display chance of winning!!

https://github.com/fohristiwhirl/nibbler

https://github.com/fohristiwhirl/nibbler/releases

Notice that a config.json file has to be created to reflect the path of your Lc0 locations.

This is analysis of Leela for initial position in my desktop!!

This is the analysis showing all possible moves!

(Note:  : She hate 1. g4 most , she feels her winning chance drops to 30%)

This is very nice.

I already mentioned the GUI link in the post.

You must have a working Leela engine in your desktop initially.

If you are new to Leela, this blog will help you set up your Leela.

http://blog.lczero.org/search?q=install+leela&m=1

Good  job, I am surprised that you can sort out by your own. You must have decent software knowledge.

Hello master, what do you think about the theme of this link (https://www.chess.com/forum/view/general/alphazero-vs-leela-zero-when-will-this-duel-happen) that I opened here on chess.com?

What about Leela's T40 and T60 networks, which one is the best today? Is T60 already stronger than T40?

T60 is already better than T60 provided that you have decent graphic card, Nividia 1060 GTX or better.

is alfa zero and leela chess zero are same engines ? if not why i did not found any games between them ?

A0 and Lc0 are bot the same engine. A0 is just an old prototype which has been abolished 2 years ago.

Lc0 is a different engine with a similar net structure.

Alpha Zero is no prototype but a highly evolved engine whose history began 2015 with Alpha Go. Unlike Lc0, Alpha Zero plays Go, shogi and chess.

A prototype means early sample. Of course A0 is early successful Neural Network of Deep Mind. A0 had only one training publicly which is 44 million games or 4 hour run ( they will definately had earlier successful versions but we never know).

Since then, there is no more improvements as they run chess as a successful example of Supervised Learning.

I have been the follower and supporter of  Leela Zero since Feb 2018  when they first release the initial program of Leela Zero in talkchess forum. ( At that time only a few dozen of people know how to run Leela Zero). I did a lot of posts how to use Leela Zero at that time in talkchess and here.

The next improvement was MuZero.

So leela chess zero vs alpha zero cannot happen?

I'd use leela if only I had a decent GPU.

You can now jun Leela on a Jetson Xavier as a server, and access it from your Windows machine.

Oh hey it's me again