Комментарии:
...but why? Why do any of this?
ОтветитьIn understand the object detection part using YOLOv5 but it seems like all it does is using NN to detect object but where did the training happen where the AI ultimately can take certain actions/decisions based on the object it detected? I want to know more *sobs*
Ответитьepic work
ОтветитьHi River, awesome vid! I'm very curious how you can control your mouse through the dongle, I've searched everywhere but they mainly point towards sniffing and malicious intent. I'm wondering if you can point me in the right direction / resources of how you achieved this. It's highly appreciated it thank you in advance!
ОтветитьWhat program do you use because i want to do the same but in a diferent game
Ответитьgive it aimbot
ОтветитьCoral USB Accelerator?
ОтветитьI want the code lol
ОтветитьHi, does this thing makes you ban at games ?. im quite interested if this concept is applicable in other games such as RPG( Genshin, honkai etc..)
ОтветитьI'm giving a wild guess and saying the reason why your fps was low is because it uses your cpu, not your gpu?
ОтветитьHow u do Mouse move?
Ответитьcan you make again a video on this ?
Ответитьai ai ai ai ai, scripting. what part of it is intelligence if you're telling it what to do?
ОтветитьReal life KAYO
Ответить@River's can you tell me what annotation software are you using?
ОтветитьNcie try! This is very interesting
ОтветитьI want to make something like this for mundane tasks in Destiny 2. Specifically fishing in Destiny 2. It’s going to be around until 2024 so might aswell Ai visual Macro Fish.
ОтветитьI don't understand why you feel the need to use another computer at all. Why not just get rid of that latency problem? Seems unnecessary.
ОтветитьHello. Loved your video! A lot going on in there. If you could explain how is the fastest way that you used to label the images, I'm not sure how to do it. And if possible, go a bit more in depth in how did you use the mouse. Anyway. Thanks.
ОтветитьI was like wow congrats you are making a bot, big deal
But guess making it only have to have inputs and output as a human player is quite interesting cuz we think that robots are better than humans because we are literally playing in them so they have full and instant control and this actually shows if humans and robot were the same who will be better?
Until the bot hosters/cheaters Crack the code and make impossible to detect cheats
Yeah and as a game developer how do you fight that kind of botting. Your model maybe is not good enough yet but it’s not that hard to make it aim perfectly. You just take bigger model and train it using simultaneous games at once. I mean it’s really ez to scale things once you have application layer ready. So now you have a bot that is undistinguishable from human playing and it aims perfectly. We are witnessing the dawn of computer gaming over internet. We will soon be going to computer clubs to play on lan just to experience PvP where each P is actual human. Who could have thought.
ОтветитьI love python
ОтветитьGPT 5 gonna have video capabilities. Imagine it doing all of that grunt work. We're teaching it to... see.
Ответитьnow i can rage quit without a ban
Ответитьno afk
ОтветитьI think if you added more info to the AI map it would do a lot better. The map is easy and clean place to get info. Like last remembered place seen, bomb site location, corners to hold etc.
Would require manual labor
Só this dude probably just was the first to actually start an bot infestation into the game
Hate people like these
is it necessary to run the bot on a seperate pc? what if its a simpler game that fps doesnt matter that much?
ОтветитьJust need to add some reinforcement learning ;)
ОтветитьYOU SHOULD DO THIS WITH OVERWATCH! IT WOULD BE INSANE!
ОтветитьI watched this video about an year ago not sure what ML really is. Learning CV now and it all makes sense. This clip got me somewhat interested in AI. Thank you!
ОтветитьYou can tune the model on a set of frames that it does not understand, this can significantly improve its quality in your task
ОтветитьYou destroy valorant 😂 company, i need this game 🤣 fucking game ban
ОтветитьWow!! Could you tell me with which sofware, you labled the photo?
Ответитьhave you heard of bobbybrocoli?
ОтветитьIt would be sick if you just let it keep learning and developing over time to see how actually good it could get.
ОтветитьWhy do you sound like Garbarj
Ответитьthis is literally the silent aim bot which means its a bit harder to get banned
ОтветитьWhat you used to send signal to the logitech dongle? Realy interested!
ОтветитьMan you just made it possible. I appreciate it . Vanguard is doomed here
ОтветитьI am super interested in training AI to play games as a human would. The idea of training one up to play Team Fortress 2 or Garry's Mod Horde mode with me and my friends is a really cool concept. I'd love to eventually try this out on my own but it's a very daunting task.
ОтветитьI say that you should give it better data. Like label twice as much. Make a auto labeler. Download pro streams of valorant, use the auto labeler and you have very good training data. Next, would be the AI. I don't know how you would handle it but you should give it as much as information as you can and unify its outputs so two bots aren't doing uncoordinated things or getting higher performance. Performance was an issue, might want to check out Qt5 was the module I am thinking and apply direct software bindings to get the Keyboard. All this third party is probably making it slower, so cut down on that. Overall, I am surprised by it's performance.
Other Improvements:
Audio
Abilities
Defusing
Player Count
Spatial Awareness
can you tell me which software is used to explain suff in this video i am taking about the white grid environment one
ОтветитьWould love to connect about a similar project. Where can I find your email?
Ответитьdo you have the dataset? i kinda wanna work on it for my AI project
ОтветитьOh, wow, these are my teammates in my rank up games
ОтветитьI'm doing a similar project. I have everything setup to get the x and y coordinates of the bounding boxes sent over to the crazy radio. I just had a question regarding that, is there any source/documentation or something for me to better understand how you did the mouse movements in relation to the coordinates? My crazy radio dongle detects the logitech usb using jackit if that help, and Im able to send some basic ducky scripts to it.
Ответить