Unity ML-Agents - 5 things you didn't know about - Version 1.0+
Вставка
- Опубліковано 9 чер 2024
- Machine Learning in Unity via ML-Agents 1.0+ is a great way to get started with the field, but it is also a huge framework with lots of things you probably don't know. This series covers the most recent version of Unity3Ds ML-Agents (Release Version 1 and newer, so don't worry - everything is up to date!)
Link to Hyperparameters for Gail: github.com/Unity-Technologies...
Support me on Patreon: www.patreon.com/user?u=25285137
Keep in touch: / sebastianschuc7 - Наука та технологія
Hello everybody! I have created a Discord Channel for everybody wanting to learn ML-Agents. It's a place where we can help each other out, ask questions, share ideas, and so on. You can join here: discord.gg/wDPWsQT
you probably dont give a shit but does anybody know of a trick to get back into an instagram account?
I stupidly forgot the login password. I would love any tricks you can offer me
@Briar Coleman instablaster :)
@Parker Mekhi Thanks for your reply. I found the site through google and I'm trying it out atm.
Looks like it's gonna take a while so I will reply here later when my account password hopefully is recovered.
@Parker Mekhi It worked and I actually got access to my account again. I'm so happy:D
Thanks so much, you saved my ass !
@Briar Coleman No problem =)
I'm working ML on my indie game and then come to this, bravo man!
Nice! I subscribed to your channel after the amazing GPT3 video!
Great content! very helpful
Amazing content bro😎
Great content! I was unware of 5 things before watching this video
This was really useful. I am keen to understand GAIL to improve the performance of my game play Agent.
It is just sad that the ML-Agents community have all gone quite and given up over the last year. This all seemed so very promising a year or so ago.
really great video! :)
Love it 😍
I would like to propose to you if you could do an ML-Agents project, doing a combat between 2 characters, who have some abilities (attack, defense), with their respective cooldowns, something similar to a 1vs1 combat from the battlerite video game, to the combat zone you can put some defensive blocks so that the characters can be protected from damage. I would like to see how you approach this, it would be an incredible thing to see
Excellent work, please make more tips videos. And tutorials :-)
Make more of these please, this has helped me so much! Where do you find out these features?
Hey! Thanks :) - Mostly by just using the Framework and reading the documentation.
Good video 🙌🙌
You are so underrated
good job
Hi
Sebastian, I have been working with ML-Agents for a week now and I have a doubt regarding this:
I want to display the Step count (which is outputted in the command line) to the game screen. So for this, I did use the value given by the academy code as [Academy.Instance.TotalStepCount] but while training this value doesn't match with the value which is outputted in cmd.
Do you have any idea why is this happening or am I accessing the wrong value?
Hi, marvelous videos and thank you. Is possible to incorporate ML Agents from Unity to any project with in real life, like a Arduino or Raspberry application?
I would say its possible for sure, but probably a lot of work and quite experimental, too. What are you planning?
@@SebastianSchuchmannAI nothing in special, maybe a content suggestion to see in future. Thanks for reading.
You know How can edit the parameters pf the neural network? Whats the file where is created it?
Hi Sebastian, I had a problem when I wanted to use the most recent version of ML-Agents, it seems that it is due to my processor, if I learn using an old version, for example 0.3, this knowledge could be applied in the same way to future more recent versions ?. I would like you to do a course linked to ML-Agents, there are few instructors who teach about this, thanks
My problem has been due to an error occurring when loading tensorflow, from what I understand, I can only run at most the version of tensorflow 1.5
Cool stuff, how about showing a basic example of an enemy AI (simple FPS) attacking the Player. In most AI examples they're pretty simple, they go towards the player and shoot. I would think AI would at least respond to being attacked, dodging fire perhaps, ducking... too complex?
James C yeah i wanna see something like this too, could be cool to actually see a more advanced ai via ml agents
Hey Sebastian ! I am trying to teach my AI to recognize sign language , do you have any tips for me ? Do you think it is possible with ml-agent ? I am using leap motion camera which makes a skeleton of the hand . Lots of love to you and thanks for the videos !
Seems more like a supervised learning task using CNNs. ML-Agents is a reinforcement learning Framework, so just Tensorflow or PyTorch is probably a better fit for you. Maybe you can find some Datasets online but I guess they will probably contain Image Data instead of skeleton Data. Of course using the data of the leap could reduce your training time but also limits the application of your model to users who have one. Very interesting problem and a great idea!
@@SebastianSchuchmannAI thanks for the tips and the reply ! Yeah we wanted to create somewhat of a sandbox type game where you can learn sign language , we wanted to try using datasets ! But didnt really know if ml-agent would come in handy . Have a nice day and cant wait for next video
The AI when penalty is too harsh 😔💥🔫
first view, first comment
Nice!