TalkChess.com

lkaufman wrote: ↑Sun Mar 31, 2019 8:07 am One question: does anyone know roughly how many 2080s would be needed to duplicate the training that this 51xxx series has averaged so far? It doesn't mean much to say that it has trained for two days without stating what the average resources used for the training were. I imagine that they were just a tiny percentage of the resources used to train AlphaZero in 9 hours or so.

Albert Silver wrote: ↑Sun Mar 31, 2019 5:42 pm

jp wrote: ↑Sun Mar 31, 2019 8:37 am NN 51010 has had 559441 games. The NN size is smaller (10x128) than the A0 one and 41xxx NNs (20x256), but the training param is visits=10000 (visits=800 for 41xxx).

That is inaccurate. It is averaging around 800 visits per move too.

Uri Blass wrote: ↑Sun Mar 31, 2019 8:27 pm

Ozymandias wrote: ↑Sun Mar 31, 2019 12:23 pm

Uri Blass wrote: ↑Sun Mar 31, 2019 11:20 amThe claim is that TCEC used some opening book to reduce the number of draws and with good opening book we are going to get almost 100% draws.

larry's words about it:
"It seems pretty clear that if you take the strongest Stockfish and the strongest NN engine on TCEC type hardware, and have them play a long match at TCEC time controls, the results will depend heavily on the openings. If you give each side a really good, deep opening book and have them play only the optimum or near optimum openings, nearly every game will end in a draw."

I am not sure about it and I have 2 problems:
1)How do you define if a book is a good deep opening book?
2)How do you define if an opening is the optimum or near the optimum opening?

I'm surprised to see this topic still being debated after so long, maybe you missed the draw death of Freestyle chess. The answer to your questions is there.

The draw death of freestyle chess proves nothing because it is possible that the participants are simply at equal level and it is possible to beat them by playing better than them.

There are also OTB tournaments when most of the games are drawn and not because the players make no mistakes.

Albert Silver wrote: ↑Sun Mar 31, 2019 6:37 pm

lkaufman wrote: ↑Sun Mar 31, 2019 6:30 pm

Laskos wrote: ↑Sun Mar 31, 2019 10:58 am

jp wrote: ↑Sun Mar 31, 2019 8:37 am

lkaufman wrote: ↑Sun Mar 31, 2019 8:07 am

NN 51010 has had 559441 games. The NN size is smaller (10x128) than the A0 one and 41xxx NNs (20x256), but the training param is visits=10000 (visits=800 for 41xxx).

I'll let others say what 2080 time that works out to.

I will try to estimate (writing on the phone, I am on a vacation ). About 0.2s/move on 2080, meaning games in about 20s. About 200 games per hour, so half a million games need 2500 hours, or about 100 days needed for training the NN 51010 on a single 2080. Or 100x 2080 GPUs needed to train it in one day.
The main unknown is the time to get 10000 visits, I took it as 0.2s on 2080. So this calculation is just for the order of magnitude.

Thanks. So the training resources were a lot more than I thought they were. It makes the amazing results at least easier to accept.

I will repeat: the average nodes per visit in this training run is around 800 nodes, not 10 thousand. It is a complete misunderstanding of what is being done. The average nodes per move per game is around 800 still.

jp wrote: ↑Sun Mar 31, 2019 9:47 pm

Albert Silver wrote: ↑Sun Mar 31, 2019 6:37 pm

lkaufman wrote: ↑Sun Mar 31, 2019 6:30 pm

Laskos wrote: ↑Sun Mar 31, 2019 10:58 am

jp wrote: ↑Sun Mar 31, 2019 8:37 am

lkaufman wrote: ↑Sun Mar 31, 2019 8:07 am

NN 51010 has had 559441 games. The NN size is smaller (10x128) than the A0 one and 41xxx NNs (20x256), but the training param is visits=10000 (visits=800 for 41xxx).

I'll let others say what 2080 time that works out to.

I will try to estimate (writing on the phone, I am on a vacation ). About 0.2s/move on 2080, meaning games in about 20s. About 200 games per hour, so half a million games need 2500 hours, or about 100 days needed for training the NN 51010 on a single 2080. Or 100x 2080 GPUs needed to train it in one day.
The main unknown is the time to get 10000 visits, I took it as 0.2s on 2080. So this calculation is just for the order of magnitude.

Thanks. So the training resources were a lot more than I thought they were. It makes the amazing results at least easier to accept.

I will repeat: the average nodes per visit in this training run is around 800 nodes, not 10 thousand. It is a complete misunderstanding of what is being done. The average nodes per move per game is around 800 still.

I will repeat: visits=10000 for NN 51010 (visits=800 for 41xxx) is what the lc0 website says.

nabildanial wrote: ↑Sun Mar 31, 2019 10:13 pm

jp wrote: ↑Sun Mar 31, 2019 9:47 pm

Albert Silver wrote: ↑Sun Mar 31, 2019 6:37 pm

lkaufman wrote: ↑Sun Mar 31, 2019 6:30 pm

Laskos wrote: ↑Sun Mar 31, 2019 10:58 am

jp wrote: ↑Sun Mar 31, 2019 8:37 am

lkaufman wrote: ↑Sun Mar 31, 2019 8:07 am

NN 51010 has had 559441 games. The NN size is smaller (10x128) than the A0 one and 41xxx NNs (20x256), but the training param is visits=10000 (visits=800 for 41xxx).

I'll let others say what 2080 time that works out to.

I will try to estimate (writing on the phone, I am on a vacation ). About 0.2s/move on 2080, meaning games in about 20s. About 200 games per hour, so half a million games need 2500 hours, or about 100 days needed for training the NN 51010 on a single 2080. Or 100x 2080 GPUs needed to train it in one day.
The main unknown is the time to get 10000 visits, I took it as 0.2s on 2080. So this calculation is just for the order of magnitude.

Thanks. So the training resources were a lot more than I thought they were. It makes the amazing results at least easier to accept.

I will repeat: the average nodes per visit in this training run is around 800 nodes, not 10 thousand. It is a complete misunderstanding of what is being done. The average nodes per move per game is around 800 still.

I will repeat: visits=10000 for NN 51010 (visits=800 for 41xxx) is what the lc0 website says.

Correct me if I'm wrong, but as far as I know starting from T50, there's another experiment that enables Lc0 to dynamically use different visit counts depending on the complexity of each positions. Visit=10000 is the upper bound, lc0 might use as low as 1 visit on a position but the average visit is still around 800 visits per game in training.

nabildanial wrote: ↑Sun Mar 31, 2019 10:17 pm Correct me if I'm wrong, but as far as I know starting from T50, there's another experiment that enables Lc0 to dynamically use different visit counts depending on the complexity of each positions. Visit=10000 is the upper bound, lc0 might use as low as 1 visit on a position but the average visit is still around 800 visits per game in training.

jp wrote: ↑Sun Mar 31, 2019 10:21 pm

nabildanial wrote: ↑Sun Mar 31, 2019 10:17 pm Correct me if I'm wrong, but as far as I know starting from T50, there's another experiment that enables Lc0 to dynamically use different visit counts depending on the complexity of each positions. Visit=10000 is the upper bound, lc0 might use as low as 1 visit on a position but the average visit is still around 800 visits per game in training.

That would make sense. And before T50 was it fixed to be exactly 800?

I'd guess there'd have to be another parameter for that.
Maybe there's something on Discord, is there?

What is "kld"?

nabildanial wrote: ↑Sun Mar 31, 2019 10:13 pm

jp wrote: ↑Sun Mar 31, 2019 9:47 pm

Albert Silver wrote: ↑Sun Mar 31, 2019 6:37 pm

lkaufman wrote: ↑Sun Mar 31, 2019 6:30 pm

Laskos wrote: ↑Sun Mar 31, 2019 10:58 am

jp wrote: ↑Sun Mar 31, 2019 8:37 am

lkaufman wrote: ↑Sun Mar 31, 2019 8:07 am

NN 51010 has had 559441 games. The NN size is smaller (10x128) than the A0 one and 41xxx NNs (20x256), but the training param is visits=10000 (visits=800 for 41xxx).

I'll let others say what 2080 time that works out to.

I will try to estimate (writing on the phone, I am on a vacation ). About 0.2s/move on 2080, meaning games in about 20s. About 200 games per hour, so half a million games need 2500 hours, or about 100 days needed for training the NN 51010 on a single 2080. Or 100x 2080 GPUs needed to train it in one day.
The main unknown is the time to get 10000 visits, I took it as 0.2s on 2080. So this calculation is just for the order of magnitude.

Thanks. So the training resources were a lot more than I thought they were. It makes the amazing results at least easier to accept.

I will repeat: the average nodes per visit in this training run is around 800 nodes, not 10 thousand. It is a complete misunderstanding of what is being done. The average nodes per move per game is around 800 still.

I will repeat: visits=10000 for NN 51010 (visits=800 for 41xxx) is what the lc0 website says.

Correct me if I'm wrong, but as far as I know starting from T50, there's another experiment that enables Lc0 to dynamically use different visit counts depending on the complexity of each positions. Visit=10000 is the upper bound, lc0 might use as low as 1 visit on a position but the average visit is still around 800 visits per game in training.

nabildanial wrote: ↑Sun Mar 31, 2019 10:35 pm Correct. You can read more about that here under "kldgain thresholding in training" topic.