Search found 101 matches
- Mon Mar 16, 2020 8:56 pm
- Forum: Computer Go
- Topic: KataGo V1.3
- Replies: 215
- Views: 132054
Re: KataGo V1.3
It depends on your strength, of course. For me, I can't even beat it with 1 playout.
- Mon Mar 09, 2020 5:40 pm
- Forum: Computer Go
- Topic: KataGo V1.3
- Replies: 215
- Views: 132054
Re: KataGo V1.3
When a bot knows ladder, many interesting things happen, like this one.
- Mon Feb 17, 2020 10:39 am
- Forum: Computer Go
- Topic: New networks Leela Zero, a Sensible Artificial Intelligence
- Replies: 87
- Views: 87239
Re: New networks Leela Zero, a Sensible Artificial Intellige
Don't forget the capability to play handicap games. Although the 40 block network is powerful, it basically has not much idea how to play handicap games.
- Thu Jan 30, 2020 6:00 pm
- Forum: Computer Go
- Topic: Elf v2 @ 800 playouts
- Replies: 13
- Views: 8557
Re: Elf v2 @ 800 playouts
Well I never heard any amateur 10 dan rank. We think we can safely put it into the professional rank range. However, it may be hard to invite any professionals to do any match with this bot. Just a wild guess, 1600p maybe around 9p.
- Tue Dec 31, 2019 9:08 pm
- Forum: Computer Go
- Topic: A new run of KataGo released - strength comparable to ELFv2
- Replies: 126
- Views: 80729
Re: A new run of KataGo released - strength comparable to EL
I have attached my configuration file and I'm happy to hear any suggestions to make it better. During that game the root visits was around 10000 to 20000, while the ELF playouts was around 2000 to 5000.
- Tue Dec 31, 2019 11:22 am
- Forum: Computer Go
- Topic: A new run of KataGo released - strength comparable to ELFv2
- Replies: 126
- Views: 80729
Re: A new run of KataGo released - strength comparable to EL
I was running a match between Katago(B) and ELF(W). Yes I used time parity with 20 seconds per move. The playouts are variable, but ELF used roughly 5000 playouts per move. It looks like ELF was ahead from the beginning to the end. I didn't set any resign threshold but at the end both sides ...
- Mon Dec 30, 2019 4:32 pm
- Forum: Computer Go
- Topic: A new run of KataGo released - strength comparable to ELFv2
- Replies: 126
- Views: 80729
Re: A new run of KataGo released - strength comparable to EL
For higher playouts (visits), Katago is still much weaker than ELFv2.
- Mon Nov 11, 2019 7:20 pm
- Forum: Computer Go
- Topic: New networks Leela Zero, a Sensible Artificial Intelligence
- Replies: 87
- Views: 87239
Re: New networks Leela Zero, a Sensible Artificial Intellige
Zen behaved weird in the last few moves, which seem to be low level blunders.
- Thu Aug 29, 2019 10:41 pm
- Forum: Computer Go
- Topic: LeelaZero adventures on Fox
- Replies: 121
- Views: 130163
Re: LeelaZero adventures on Fox
So what is the explanation for this straight win with only hundred or two playouts? No strong opponent yet?
- Sat Aug 17, 2019 12:19 pm
- Forum: Computer Go
- Topic: A new run of KataGo released - strength comparable to ELFv2
- Replies: 126
- Views: 80729
Re: A new run of KataGo released - strength comparable to EL
It is very hard to beat ELF when it plays white.
- Wed Jul 17, 2019 7:53 pm
- Forum: Computer Go
- Topic: A new run of KataGo released - strength comparable to ELFv2
- Replies: 126
- Views: 80729
Re: A new run of KataGo released - strength comparable to EL
The most confusing part is to compile the cudahelpers.cu into an obj file. I made a batch file to run in command prompt with this contents:
NVCC cudahelpers.cu -c --cudart shared -gencode arch=compute_30,code=sm_30 -gencode arch=compute_30,code=compute_30 -gencode arch=compute_37,code=sm_37 ...
NVCC cudahelpers.cu -c --cudart shared -gencode arch=compute_30,code=sm_30 -gencode arch=compute_30,code=compute_30 -gencode arch=compute_37,code=sm_37 ...
- Mon Jul 01, 2019 5:57 pm
- Forum: Computer Go
- Topic: A new run of KataGo released - strength comparable to ELFv2
- Replies: 126
- Views: 80729
Re: A new run of KataGo released - strength comparable to EL
You are right it is a complex situation. However, Katago thought b was losing (w can play A5 to make it a direct ko). That's a little confusing to me. By the way, white just need to find a small ko threat to win the game. I would think it would be better than sitting there and doing nothing about ...
- Mon Jul 01, 2019 12:35 pm
- Forum: Computer Go
- Topic: A new run of KataGo released - strength comparable to ELFv2
- Replies: 126
- Views: 80729
Re: A new run of KataGo released - strength comparable to EL
Now this is a strange game between Katago(B) and LZ230(W). Near the end of the game white played F8 so that there was a seki in ko situation.
https://hosting.photobucket.com/images/h469/SPLee999/0/7a769260-392d-4cd8-8045-7923a08bec25-original.png?width=1920&height=1080&fit=bounds
Both sides ...
https://hosting.photobucket.com/images/h469/SPLee999/0/7a769260-392d-4cd8-8045-7923a08bec25-original.png?width=1920&height=1080&fit=bounds
Both sides ...
- Sat Jun 22, 2019 1:34 pm
- Forum: Computer Go
- Topic: A new run of KataGo released - strength comparable to ELFv2
- Replies: 126
- Views: 80729
Re: A new run of KataGo released - strength comparable to EL
Here is a 4 stone handicap game with Leela 0.11 with 20 second per move time setting. With the same time setting Leela Zero 40b weights have only small chance to win in 3 handicap games. The obvious style of Katago is that it likes to cut and deal with multiple groups properly. It even outsmarts ...
- Fri Jun 21, 2019 5:40 pm
- Forum: Computer Go
- Topic: A new run of KataGo released - strength comparable to ELFv2
- Replies: 126
- Views: 80729
Re: A new run of KataGo released - strength comparable to EL
Compared to ELFv2, Katago's handicap game style is really the best. It can play 5 handicap games ( I haven't test more handicaps yet) without any bad moves (remember ELF's ladder moves?).