I have an Rtx 2060 6 GB. It may be because of the lc0 network I was using. I think depending on the size of the network you may get faster nodes per second. Thanks for the comment!
@@zenchess thanks for the quick reply :) I was running it on my Thinkpad and it was nearly not as fast as yours. It makes sense that depending on the size of the network the node generation is faster. I'll try different network files.
@@asdfasdfasdf383 It could also be the GPU depending on what you have in your laptop. RTX is where leela really started taking off due to ai cores or something like that. By the way you can use stockfish in nibbler, you will just be limited in the number of move suggestions you get (I think somewhere around 3 arrows drawn and not all the features of nibbler but still easier to use than raw variations in chessbase). Stockfish nowadays also uses a fast neural network architecture and has kind of taken over as the king of engines (and runs fast on cpu so will probably perform better on your hardware)
@Zenchess I've just downloaded Nibbler today. So far it seems like a wonderful piece of software. It just works and it's easy to use. Yes, Stockfish on CPU would probably be quite strong on my machine. I downloaded and built Leela because of the more human-like, "creative" move suggestions, as opposed to traditional chess engines. That's quite interesting and I want to explore that. Stockfish with neural networks (NNUE) was introduced in 2018, and has surpassed the previous versions by a large margin. I seems to be using a kind of hybrid approach.
You get quicker visual feedback of moves in nibbler than you do in any normal gui. You can even use it with stockfish, but it will only show the top 3 moves I believe. Leela is also a very strong engine, even though it is weaker than stockfish in some positions it is actually stronger.
@@zenchess If it's stronger in some positions then why is stockfish always rated higher? And does that mean Leela has the potential to be an overall better engine than Stockfish?
I think serious chess professionals will use both leela and stockfish to evaluate a position. Stockfish is in general stronger but leela can be stronger in certain positions. However when analyzing a game or exploring chess like I showed it really doesn't matter all that much which engine you're using if they're high -rated engines like stockfish and leela. The main benefit of the nibbler gui is you get very quick feedback on the actual board of what the best moves are so you can very quickly learn the positions. You could do the same with stockfish loaded in nibbler you would just see the top 3 moves instead of all the extra moves with leela. I have no idea what the future will hold for leela vs stockfish, I think stockfish is ahead in the race but you never know what will happen. There were times when I think leela was actually stronger but it might have been a few years. The reason an engine can be stronger in a specific position than a higher rated engine is because it may have a better understanding of certain positions, but in general calculate worse. In the early days of neural network engines they might play very strong but miss a mate in 4 because they didn't calculate so well.
Why are you only getting 5k nodes per second? I get 35 million per second. And in 1 minute, I already have 1 billion
No idea. Could be because of the different network. Keep in mind this video was from 3 years ago
@@zenchessWhat do you call network?
@@NACHO55G The leela network weights. You're running leela right? It doesn't calculate as fast as stockfish
@@zenchessNo, Stockfish 16
@@zenchessDo you have any idea how to make Stockfish 16 continue calculating while waiting for the user to make a move?
Great Video
The engine lc0 is generating nodes rather quickly. Do you have a high-performance GPU?
I have an Rtx 2060 6 GB. It may be because of the lc0 network I was using. I think depending on the size of the network you may get faster nodes per second. Thanks for the comment!
@@zenchess thanks for the quick reply :) I was running it on my Thinkpad and it was nearly not as fast as yours.
It makes sense that depending on the size of the network the node generation is faster. I'll try different network files.
@@asdfasdfasdf383 It could also be the GPU depending on what you have in your laptop. RTX is where leela really started taking off due to ai cores or something like that. By the way you can use stockfish in nibbler, you will just be limited in the number of move suggestions you get (I think somewhere around 3 arrows drawn and not all the features of nibbler but still easier to use than raw variations in chessbase). Stockfish nowadays also uses a fast neural network architecture and has kind of taken over as the king of engines (and runs fast on cpu so will probably perform better on your hardware)
@Zenchess I've just downloaded Nibbler today. So far it seems like a wonderful piece of software. It just works and it's easy to use.
Yes, Stockfish on CPU would probably be quite strong on my machine. I downloaded and built Leela because of the more human-like, "creative" move suggestions, as opposed to traditional chess engines. That's quite interesting and I want to explore that.
Stockfish with neural networks (NNUE) was introduced in 2018, and has surpassed the previous versions by a large margin. I seems to be using a kind of hybrid approach.
great video
thanks!
Whats the point of using this when stockfish exists though?
You get quicker visual feedback of moves in nibbler than you do in any normal gui. You can even use it with stockfish, but it will only show the top 3 moves I believe.
Leela is also a very strong engine, even though it is weaker than stockfish in some positions it is actually stronger.
@@zenchess If it's stronger in some positions then why is stockfish always rated higher? And does that mean Leela has the potential to be an overall better engine than Stockfish?
I think serious chess professionals will use both leela and stockfish to evaluate a position. Stockfish is in general stronger but leela can be stronger in certain positions. However when analyzing a game or exploring chess like I showed it really doesn't matter all that much which engine you're using if they're high -rated engines like stockfish and leela. The main benefit of the nibbler gui is you get very quick feedback on the actual board of what the best moves are so you can very quickly learn the positions. You could do the same with stockfish loaded in nibbler you would just see the top 3 moves instead of all the extra moves with leela.
I have no idea what the future will hold for leela vs stockfish, I think stockfish is ahead in the race but you never know what will happen. There were times when I think leela was actually stronger but it might have been a few years.
The reason an engine can be stronger in a specific position than a higher rated engine is because it may have a better understanding of certain positions, but in general calculate worse. In the early days of neural network engines they might play very strong but miss a mate in 4 because they didn't calculate so well.