Search found 143 matches

by jkiliani
Sun May 27, 2018 9:07 am
Forum: Computer Chess Club: General Topics
Topic: 20x256 Leela seemed quite strong
Replies: 47
Views: 8557

Re: 20x256 Leela seemed quite strong

Virtual loss bug is a (different) type of FPU reduction, one that prunes little at the root but a lot at the leaves. Kind of strange to combine negative regular FPU reduction with positive FPU reduction by virtual loss, but if it works for you... So are you saying FPU prunes a lot at the root & lit...
by jkiliani
Sun May 27, 2018 8:34 am
Forum: Computer Chess Club: General Topics
Topic: 20x256 Leela seemed quite strong
Replies: 47
Views: 8557

Re: 20x256 Leela seemed quite strong

I didn't fiddle with scale thinking time, just took the CLOP result and common sense in time usage. I put it in the description of settings to show the difference from the default. Virtual loss bug is affecting a bit the tactical performance, although I don't even know what it stands for. I know on...
by jkiliani
Wed May 23, 2018 5:12 pm
Forum: Computer Chess Club: General Topics
Topic: lc0-win-20180512-cuda90-cudnn712-00
Replies: 104
Views: 12442

Re: lc0-win-20180512-cuda90-cudnn712-00

For self-play, ideally you don't care at all and run everything serial, and then batch over concurrent games. This gives you the best of both worlds: full batching speedup and no loss of search efficiency. But I don't think Leela Chess Zero has this capability in the client or engines, and AFAIK no...
by jkiliani
Tue May 22, 2018 7:49 pm
Forum: Computer Chess Club: General Topics
Topic: LCZero: Progress and Scaling. Relation to CCRL Elo
Replies: 604
Views: 104417

Re: LCZero: Progress and Scaling. Relation to CCRL Elo

At TC like 15m + 15s, the CCRL rating is even higher, maybe 3250 or so. The snag is that these later nets scale not that well with TC (or playouts) compared to NN237. You might try Id 329 sometime. A test of the value head today yielded the best result of any network so far, for 329, which is a ver...
by jkiliani
Sun May 20, 2018 9:44 pm
Forum: Computer Chess Club: General Topics
Topic: LCZero: Progress and Scaling. Relation to CCRL Elo
Replies: 604
Views: 104417

Re: LCZero: Progress and Scaling. Relation to CCRL Elo

Are you sure that at low node count condition the new networks perform better but scale worse is not because of over-fitting? The value head of recent nets used to be still very slight worse than that of 237. With 321, this may not be the case anymore according to some measurements. The reduced qua...
by jkiliani
Sun May 20, 2018 8:14 pm
Forum: Computer Chess Club: General Topics
Topic: LCZero: Progress and Scaling. Relation to CCRL Elo
Replies: 604
Views: 104417

Re: LCZero: Progress and Scaling. Relation to CCRL Elo

It seems to be a bit more complicated. At 1'+ 1'' on GTX 1060 with LC0 CUDA, latest nets seem even stronger than NN237. But at 15'+ 15'', NN237 seems stronger. I left NN319 against Komodo 10.2, it lost 5 games in a row due to tactical blunders. Eval graph was also unstable. I interrupted the match ...
by jkiliani
Sat May 19, 2018 12:20 am
Forum: Computer Chess Club: General Topics
Topic: LCZero: Progress and Scaling. Relation to CCRL Elo
Replies: 604
Views: 104417

Re: LCZero: Progress and Scaling. Relation to CCRL Elo

Greetings, any update on the possibility of rollback? Now that the network is steadily recovering is the rollback still being considered, or is there already some final decision on that issue? Assuming the new nets will soon reach the old nets peak, is there any advantage in doing rollback instead ...
by jkiliani
Fri May 18, 2018 9:58 am
Forum: Computer Chess Club: General Topics
Topic: LCZero: Progress and Scaling. Relation to CCRL Elo
Replies: 604
Views: 104417

Re: LCZero: Progress and Scaling. Relation to CCRL Elo

The PUCT change pull request was made by Gian-Carlo Pascutto (gcp), the one who made the original Leela Zero, that Leela Chess Zero is forked from. Is gcp a random person? Plus the PUCT change was tested a lot of times to prove its worth and it was proven that it gives more Elo. Also most of the de...
by jkiliani
Thu May 17, 2018 7:51 pm
Forum: Computer Chess Club: General Topics
Topic: LCZero: Progress and Scaling. Relation to CCRL Elo
Replies: 604
Views: 104417

Re: LCZero: Progress and Scaling. Relation to CCRL Elo

Reverting to an earlier version just because something went wrong is not smart. You're actually right - letting random people commit random stuff and changing random constants without testing is not smat but rather incompetent. Problems have been pointed out by Ronald, Alvaro and other people who a...
by jkiliani
Thu May 17, 2018 7:27 pm
Forum: Computer Chess Club: General Topics
Topic: LCZero: Progress and Scaling. Relation to CCRL Elo
Replies: 604
Views: 104417

Re: LCZero: Progress and Scaling. Relation to CCRL Elo

I plan to play 200 games to get a rough idea of how strong the current engine + net is, I'll post the results here. I have to agree with Albert - after 100 games perf is -60 elo compared to ID 24x so fanboys can cry now, they should revert to 0.7 ASAP! Reverting to an earlier version just because s...