elite*gold: 0 
Registriert seit: Mar 2021
Beiträge: 870
Erhaltene Thanks: 216
|
To go more in depth then the above people.
Majority of AI/Neural Networking cheats are based upon Lunar because it takes time and resources to train a network from absolute scratch.
Lunar isn't bad in my opinion, but the creator from lunar hates it, and has started the development and training of Neura, which is 10 times better, Lunar isn't perfect no, more on that below.
If the neural network used for the cheat, is trained very very well, then it is pretty much extremely accurate, but if it isn't trained well, it can still be accurate for majority of the time however it can cause very sus situations in-game from time to time resulting in people calling you out for cheating.
And keep in mind, training such a network costs not only time, but also resources to do properly, something that majority developers don't want to do because Valorant already has clear outlines around their characters, an simple solution for them still wanting to offer AI out of marketing perspective is to make use of Lunar, but again, Lunar isn't perfect out of the box and still needs an extend of training and tweaking as it's originally created for Fortnite not valorant, but it is good enough tho with some tweaks and minor training.
To answer your question now, Neural networking (ai) is better if the characters in that game doesn't have an outline and you still want to go external, AI is your way to go then.
In valorant's case there are characters with outlines, so at first glance there isn't much of a difference, both are solid options, UNLESS the neural network is trained very very well, then it will beat color for sure, but that also depends on multiple factors again.
Also devs tend to be lazy not only in terms of going for lunar instead of from scratch but also in terms of how they train any neural network, when training a neural network, the correct way is to train it for full body, and each body part separately however most train it for full body only to save a A LOT of time, effort and resources, this i call a poorly trained ai, and color is in any situation better as when the body is half blocked, it won't be properly recognized often, while color would still recognize the color.
Is AI worth paying extra for compared to color, in valorant?
Not really no, unless it's just a couple bux extra and their neural network is trained very very well, then sure why not right.
Is it day and night difference compared to color?
In valorant definitely not no, but again that depends on how well trained the neural network in specific is, and what color recognition methods are used, does it support multi color, does it automatically recognize darker/brighter areas, etc..
Side not, you also need an nvidia gpu, a decent one, to run cuda based neural network/ai bots.
--------------------------
Skyrant is based on colors, and we've been doing that for 2-3 years now.
Recently we've started the development and training of our own Neural Network.
But not because AI is better then color, Skyrant will not replace the color detection, it will offer AI as an optional feature of character recognition.
The reason we've started on our own Neural network for Skyrant, is not purely for Valorant as color works very well in this game, but mainly to expand towards other games where character outlines are not optimal nor present at all.
|
