Same here, from 59.0% to 56.5%, a loss of almost 20 elo.Laskos wrote: ↑Thu Jul 23, 2020 5:59 amI get a strong regression for 20200723-1134.binRebel wrote: ↑Wed Jul 22, 2020 9:36 pm The last (2210) finally 59%, +63 elo above the official stockfish 11.
Code: Select all
SF-NNUE (popcount) vs Stockfish 11 henk-2706 54.7% 2020-07-19 sergio-1432 56.1% 2020-07-21 15:56 sergio-1907 58.9% 2020-07-21 19:14 sergio-2323 58.4% 2020-07-21 23:23 sergio-0359 58.8% 2020-07-22 03:59 sergio-0944 58.9% 2020-07-22 09:44 sergio-1153 58.3% 2020-07-22 11:53 sergio-1807 58.0% 2020-07-22 18:07 sergio-2210 59.0% 2020-07-22 22:10
Sergio Vieri second net is out
Moderators: hgm, Rebel, chrisw
-
- Posts: 6997
- Joined: Thu Aug 18, 2011 12:04 pm
- Full name: Ed Schröder
Re: Sergio Vieri second net is out
90% of coding is debugging, the other 10% is writing bugs.
-
- Posts: 4319
- Joined: Tue Apr 03, 2012 4:28 pm
Re: Sergio Vieri second net is out
And nobody has the slightest idea why. The whole thing is kind of ridiculous.Rebel wrote: ↑Thu Jul 23, 2020 11:39 amSame here, from 59.0% to 56.5%, a loss of almost 20 elo.Laskos wrote: ↑Thu Jul 23, 2020 5:59 amI get a strong regression for 20200723-1134.binRebel wrote: ↑Wed Jul 22, 2020 9:36 pm The last (2210) finally 59%, +63 elo above the official stockfish 11.
Code: Select all
SF-NNUE (popcount) vs Stockfish 11 henk-2706 54.7% 2020-07-19 sergio-1432 56.1% 2020-07-21 15:56 sergio-1907 58.9% 2020-07-21 19:14 sergio-2323 58.4% 2020-07-21 23:23 sergio-0359 58.8% 2020-07-22 03:59 sergio-0944 58.9% 2020-07-22 09:44 sergio-1153 58.3% 2020-07-22 11:53 sergio-1807 58.0% 2020-07-22 18:07 sergio-2210 59.0% 2020-07-22 22:10
-
- Posts: 4190
- Joined: Wed Nov 25, 2009 1:47 am
Re: Sergio Vieri second net is out
The only thing that is rediculous is your knowledge about DNN training.chrisw wrote: ↑Thu Jul 23, 2020 1:36 pmAnd nobody has the slightest idea why. The whole thing is kind of ridiculous.Rebel wrote: ↑Thu Jul 23, 2020 11:39 amSame here, from 59.0% to 56.5%, a loss of almost 20 elo.Laskos wrote: ↑Thu Jul 23, 2020 5:59 amI get a strong regression for 20200723-1134.binRebel wrote: ↑Wed Jul 22, 2020 9:36 pm The last (2210) finally 59%, +63 elo above the official stockfish 11.
Code: Select all
SF-NNUE (popcount) vs Stockfish 11 henk-2706 54.7% 2020-07-19 sergio-1432 56.1% 2020-07-21 15:56 sergio-1907 58.9% 2020-07-21 19:14 sergio-2323 58.4% 2020-07-21 23:23 sergio-0359 58.8% 2020-07-22 03:59 sergio-0944 58.9% 2020-07-22 09:44 sergio-1153 58.3% 2020-07-22 11:53 sergio-1807 58.0% 2020-07-22 18:07 sergio-2210 59.0% 2020-07-22 22:10
The loss happens because net is overfit. Problem with training of NNs as eval functions is that you can't just do validation as you do it with regular DNN training. A typical loss function is useless. You get smaller loss on your validation dataset but once you play games you get much larger loss. And there is nothing you can do about this. Coz eval is coupled with search so you can't just validate eval standalone.
-
- Posts: 4319
- Joined: Tue Apr 03, 2012 4:28 pm
Re: Sergio Vieri second net is out
Exactly what I meant, you show it. Typical “engineer” who knows nothing. You cannot explain in CHESS terms what it knows , doesnt know, or how or how not. Clueless nerd.Milos wrote: ↑Thu Jul 23, 2020 2:36 pmThe only thing that is rediculous is your knowledge about DNN training.chrisw wrote: ↑Thu Jul 23, 2020 1:36 pmAnd nobody has the slightest idea why. The whole thing is kind of ridiculous.Rebel wrote: ↑Thu Jul 23, 2020 11:39 amSame here, from 59.0% to 56.5%, a loss of almost 20 elo.Laskos wrote: ↑Thu Jul 23, 2020 5:59 amI get a strong regression for 20200723-1134.binRebel wrote: ↑Wed Jul 22, 2020 9:36 pm The last (2210) finally 59%, +63 elo above the official stockfish 11.
Code: Select all
SF-NNUE (popcount) vs Stockfish 11 henk-2706 54.7% 2020-07-19 sergio-1432 56.1% 2020-07-21 15:56 sergio-1907 58.9% 2020-07-21 19:14 sergio-2323 58.4% 2020-07-21 23:23 sergio-0359 58.8% 2020-07-22 03:59 sergio-0944 58.9% 2020-07-22 09:44 sergio-1153 58.3% 2020-07-22 11:53 sergio-1807 58.0% 2020-07-22 18:07 sergio-2210 59.0% 2020-07-22 22:10
The loss happens because net is overfit. Problem with training of NNs as eval functions is that you can't just do validation as you do it with regular DNN training. A typical loss function is useless. You get smaller loss on your validation dataset but once you play games you get much larger loss. And there is nothing you can do about this. Coz eval is coupled with search so you can't just validate eval standalone.
-
- Posts: 4190
- Joined: Wed Nov 25, 2009 1:47 am
Re: Sergio Vieri second net is out
Lol chess terms. In the dictionary under demagogue they should have your name .chrisw wrote: ↑Thu Jul 23, 2020 3:59 pmExactly what I meant, you show it. Typical “engineer” who knows nothing. You cannot explain in CHESS terms what it knows , doesnt know, or how or how not. Clueless nerd.Milos wrote: ↑Thu Jul 23, 2020 2:36 pmThe only thing that is rediculous is your knowledge about DNN training.chrisw wrote: ↑Thu Jul 23, 2020 1:36 pmAnd nobody has the slightest idea why. The whole thing is kind of ridiculous.Rebel wrote: ↑Thu Jul 23, 2020 11:39 amSame here, from 59.0% to 56.5%, a loss of almost 20 elo.Laskos wrote: ↑Thu Jul 23, 2020 5:59 amI get a strong regression for 20200723-1134.binRebel wrote: ↑Wed Jul 22, 2020 9:36 pm The last (2210) finally 59%, +63 elo above the official stockfish 11.
Code: Select all
SF-NNUE (popcount) vs Stockfish 11 henk-2706 54.7% 2020-07-19 sergio-1432 56.1% 2020-07-21 15:56 sergio-1907 58.9% 2020-07-21 19:14 sergio-2323 58.4% 2020-07-21 23:23 sergio-0359 58.8% 2020-07-22 03:59 sergio-0944 58.9% 2020-07-22 09:44 sergio-1153 58.3% 2020-07-22 11:53 sergio-1807 58.0% 2020-07-22 18:07 sergio-2210 59.0% 2020-07-22 22:10
The loss happens because net is overfit. Problem with training of NNs as eval functions is that you can't just do validation as you do it with regular DNN training. A typical loss function is useless. You get smaller loss on your validation dataset but once you play games you get much larger loss. And there is nothing you can do about this. Coz eval is coupled with search so you can't just validate eval standalone.
It's not shameful to be old and not understand new technology. But it's embarrassing trying to present it as some high level philosophy in the same time just because you don't understand it.
-
- Posts: 1631
- Joined: Tue Aug 21, 2018 7:52 pm
- Full name: Dietrich Kappe
Re: Sergio Vieri second net is out
Is the price you pay for freedom from bean counting.
Eavesdropping on discord;
1. These nets are trained 50% on score and 50% on game result (that can be adjusted)
2. The data generation inserts some random moves into the game
3. Apparently SV increased the maximum ply where this happens, causing the regression
4. Bean counters everywhere rejoiced
Fat Titz by Stockfish, the engine with the bodaciously big net. Remember: size matters. If you want to learn more about this engine just google for "Fat Titz".
-
- Posts: 4319
- Joined: Tue Apr 03, 2012 4:28 pm
Re: Sergio Vieri second net is out
Milos, neural net software is not exactly rocket science. Presenting as “overfitting” is no answer to anything, it’s a standard science-blinding which probably doesn’t even apply here, the Sergio nets are being constantly presented with new training data. Try getting your fundamentals straight before launching drama queen. Difficult and probably a bit too deep for for you, but lesson one in big boys world is not everybody is an idiot just like you were yesterday.Milos wrote: ↑Thu Jul 23, 2020 4:26 pmLol chess terms. In the dictionary under demagogue they should have your name .chrisw wrote: ↑Thu Jul 23, 2020 3:59 pmExactly what I meant, you show it. Typical “engineer” who knows nothing. You cannot explain in CHESS terms what it knows , doesnt know, or how or how not. Clueless nerd.Milos wrote: ↑Thu Jul 23, 2020 2:36 pmThe only thing that is rediculous is your knowledge about DNN training.chrisw wrote: ↑Thu Jul 23, 2020 1:36 pmAnd nobody has the slightest idea why. The whole thing is kind of ridiculous.Rebel wrote: ↑Thu Jul 23, 2020 11:39 amSame here, from 59.0% to 56.5%, a loss of almost 20 elo.Laskos wrote: ↑Thu Jul 23, 2020 5:59 amI get a strong regression for 20200723-1134.binRebel wrote: ↑Wed Jul 22, 2020 9:36 pm The last (2210) finally 59%, +63 elo above the official stockfish 11.
Code: Select all
SF-NNUE (popcount) vs Stockfish 11 henk-2706 54.7% 2020-07-19 sergio-1432 56.1% 2020-07-21 15:56 sergio-1907 58.9% 2020-07-21 19:14 sergio-2323 58.4% 2020-07-21 23:23 sergio-0359 58.8% 2020-07-22 03:59 sergio-0944 58.9% 2020-07-22 09:44 sergio-1153 58.3% 2020-07-22 11:53 sergio-1807 58.0% 2020-07-22 18:07 sergio-2210 59.0% 2020-07-22 22:10
The loss happens because net is overfit. Problem with training of NNs as eval functions is that you can't just do validation as you do it with regular DNN training. A typical loss function is useless. You get smaller loss on your validation dataset but once you play games you get much larger loss. And there is nothing you can do about this. Coz eval is coupled with search so you can't just validate eval standalone.
It's not shameful to be old and not understand new technology. But it's embarrassing trying to present it as some high level philosophy in the same time just because you don't understand it.
-
- Posts: 4190
- Joined: Wed Nov 25, 2009 1:47 am
Re: Sergio Vieri second net is out
Sorry but to have a meaningful discussion one needs to understand at least the basic concept how DNN training is done. You clearly don't understand it and are showing nothing but Dunning-Kruger here. But that's nothing new for you anyway.chrisw wrote: ↑Thu Jul 23, 2020 4:45 pmMilos, neural net software is not exactly rocket science. Presenting as “overfitting” is no answer to anything, it’s a standard science-blinding which probably doesn’t even apply here, the Sergio nets are being constantly presented with new training data. Try getting your fundamentals straight before launching drama queen. Difficult and probably a bit too deep for for you, but lesson one in big boys world is not everybody is an idiot just like you were yesterday.Milos wrote: ↑Thu Jul 23, 2020 4:26 pmLol chess terms. In the dictionary under demagogue they should have your name .chrisw wrote: ↑Thu Jul 23, 2020 3:59 pmExactly what I meant, you show it. Typical “engineer” who knows nothing. You cannot explain in CHESS terms what it knows , doesnt know, or how or how not. Clueless nerd.Milos wrote: ↑Thu Jul 23, 2020 2:36 pmThe only thing that is rediculous is your knowledge about DNN training.chrisw wrote: ↑Thu Jul 23, 2020 1:36 pmAnd nobody has the slightest idea why. The whole thing is kind of ridiculous.Rebel wrote: ↑Thu Jul 23, 2020 11:39 amSame here, from 59.0% to 56.5%, a loss of almost 20 elo.Laskos wrote: ↑Thu Jul 23, 2020 5:59 amI get a strong regression for 20200723-1134.binRebel wrote: ↑Wed Jul 22, 2020 9:36 pm The last (2210) finally 59%, +63 elo above the official stockfish 11.
Code: Select all
SF-NNUE (popcount) vs Stockfish 11 henk-2706 54.7% 2020-07-19 sergio-1432 56.1% 2020-07-21 15:56 sergio-1907 58.9% 2020-07-21 19:14 sergio-2323 58.4% 2020-07-21 23:23 sergio-0359 58.8% 2020-07-22 03:59 sergio-0944 58.9% 2020-07-22 09:44 sergio-1153 58.3% 2020-07-22 11:53 sergio-1807 58.0% 2020-07-22 18:07 sergio-2210 59.0% 2020-07-22 22:10
The loss happens because net is overfit. Problem with training of NNs as eval functions is that you can't just do validation as you do it with regular DNN training. A typical loss function is useless. You get smaller loss on your validation dataset but once you play games you get much larger loss. And there is nothing you can do about this. Coz eval is coupled with search so you can't just validate eval standalone.
It's not shameful to be old and not understand new technology. But it's embarrassing trying to present it as some high level philosophy in the same time just because you don't understand it.
-
- Posts: 4319
- Joined: Tue Apr 03, 2012 4:28 pm
Re: Sergio Vieri second net is out
Hmmm. When I need lessons from the little prince who never made it, I'll come and ask. Meanwhile, fault of the arrogant immature is to imagine that what you learnt yesterday is a new discovery that nobody else knows. Engineering ignoramus.Milos wrote: ↑Thu Jul 23, 2020 5:07 pmSorry but to have a meaningful discussion one needs to understand at least the basic concept how DNN training is done. You clearly don't understand it and are showing nothing but Dunning-Kruger here. But that's nothing new for you anyway.chrisw wrote: ↑Thu Jul 23, 2020 4:45 pmMilos, neural net software is not exactly rocket science. Presenting as “overfitting” is no answer to anything, it’s a standard science-blinding which probably doesn’t even apply here, the Sergio nets are being constantly presented with new training data. Try getting your fundamentals straight before launching drama queen. Difficult and probably a bit too deep for for you, but lesson one in big boys world is not everybody is an idiot just like you were yesterday.Milos wrote: ↑Thu Jul 23, 2020 4:26 pmLol chess terms. In the dictionary under demagogue they should have your name .chrisw wrote: ↑Thu Jul 23, 2020 3:59 pmExactly what I meant, you show it. Typical “engineer” who knows nothing. You cannot explain in CHESS terms what it knows , doesnt know, or how or how not. Clueless nerd.Milos wrote: ↑Thu Jul 23, 2020 2:36 pmThe only thing that is rediculous is your knowledge about DNN training.chrisw wrote: ↑Thu Jul 23, 2020 1:36 pmAnd nobody has the slightest idea why. The whole thing is kind of ridiculous.Rebel wrote: ↑Thu Jul 23, 2020 11:39 amSame here, from 59.0% to 56.5%, a loss of almost 20 elo.Laskos wrote: ↑Thu Jul 23, 2020 5:59 amI get a strong regression for 20200723-1134.binRebel wrote: ↑Wed Jul 22, 2020 9:36 pm The last (2210) finally 59%, +63 elo above the official stockfish 11.
Code: Select all
SF-NNUE (popcount) vs Stockfish 11 henk-2706 54.7% 2020-07-19 sergio-1432 56.1% 2020-07-21 15:56 sergio-1907 58.9% 2020-07-21 19:14 sergio-2323 58.4% 2020-07-21 23:23 sergio-0359 58.8% 2020-07-22 03:59 sergio-0944 58.9% 2020-07-22 09:44 sergio-1153 58.3% 2020-07-22 11:53 sergio-1807 58.0% 2020-07-22 18:07 sergio-2210 59.0% 2020-07-22 22:10
The loss happens because net is overfit. Problem with training of NNs as eval functions is that you can't just do validation as you do it with regular DNN training. A typical loss function is useless. You get smaller loss on your validation dataset but once you play games you get much larger loss. And there is nothing you can do about this. Coz eval is coupled with search so you can't just validate eval standalone.
It's not shameful to be old and not understand new technology. But it's embarrassing trying to present it as some high level philosophy in the same time just because you don't understand it.
-
- Posts: 1631
- Joined: Tue Aug 21, 2018 7:52 pm
- Full name: Dietrich Kappe
Re: Sergio Vieri second net is out
If talkchess had a fragrance it would be called “Bile.”
Fat Titz by Stockfish, the engine with the bodaciously big net. Remember: size matters. If you want to learn more about this engine just google for "Fat Titz".