Train your first A.I. in Unity | ML-Agents Tutorial 2020

[@BotAcademyYT]

This is the first video in my ML Agents playlist. Make sure to check out the other ML Agents videos in the playlist: kzbin.info/aero/PL8fePt58xRPY1-pkhMPus3GlUGXNdqMH5

[@shepherdaiden7194]

I know Im asking the wrong place but does anyone know a tool to get back into an instagram account?? I was stupid forgot the account password. I would love any assistance you can offer me

[@jaxonheath3038]

@Shepherd Aiden Instablaster =)

[@shepherdaiden7194]

@Jaxon Heath Thanks for your reply. I got to the site on google and Im waiting for the hacking stuff now. Seems to take quite some time so I will get back to you later with my results.

[@shepherdaiden7194]

@Jaxon Heath It worked and I finally got access to my account again. I am so happy:D Thank you so much you really help me out!

[@jaxonheath3038]

@Shepherd Aiden glad I could help :)

[@anicsim8390]

even in 2024 this video is still helping me train my agents hahah good stuff and great tutorial sir

[@wsalevan]

I've looked through so many tutorials and this one is the only one that worked. Thank you so much!

[@aniketroy7292]

Thank you very much Sir. I had been trying to achieve this for days. I read their documentation on ml agents but nothing helped much. I followed the same way u showed in this tutorial and everything worked so fine. You don't know how much you have helped!! Thanks Again :)

[@Lx30]

Took the better part of the day playing with all the pip installs and correct versions, but it all works now. Thanks for this :) good job!

[@BotAcademyYT]

Thanks! Great you got it working. Enjoy!

[@jihadalakl7964]

Nice series man, I would like to learn more and Finnish the full series. Keep going.

[@Zorfi]

yeah!

[@kirill_spirit_releasement]

Very much thank you! Your video helped me go this on my comp. although I don't even know English, everything was clear. Thank you very much!

[@ewwkl7279]

Your tutorial is so clear and well instructed. It's my first mlagents project that worked. Thank you very much.

[@BotAcademyYT]

Glad it was helpful!

[@ankusharora8082]

The best video to start working on ML-Agents and also support system is very helpful.

[@xingranruan708]

Thanks! This is the best video I have seen so far!!! I will go through all the other videos!

[@BotAcademyYT]

Thanks :-)

[@aryan_kode]

nice work please continue the series. looking forward to it

[@Mhmoudko]

This is the Most Clear Video on ML-Agents ... Thanks a lot

[@sadiqabbaszade4789]

This is my first time trying to learn ML (though i am a game dev), and I hope I can learn something from your tutorials. Demos in intro look very cool!

[@BotAcademyYT]

Thanks! Enjoy the videos :)

[@rainshih7542]

The clearest video tut so far, thanks man! looking forward to your video update.

[@BotAcademyYT]

Thank you :) Make sure to check out my second ML-Agents video if you haven’t yet. Next video should be finished by the end of this week.

[@planetdilien2932]

Please continue the series. I would love to know how to make games for it on its own!

[@blue3566]

That's exactly what I need for my next game. Thx for sharing

[@adamjurik5442]

I highly appreciate it. I had an issue using newer versions of the ML Agents and now i'm using v1, and no longer get the port error (I tried to fix this in many ways but downgrading was the only choice). Thank you!

[@BotAcademyYT]

You’re welcome. Glad I could help :)

[@annekatrinwolfrum3461]

Thank you guy i find A.I is the collest thing in Unity so thanks for this tutorial

[@BotAcademyYT]

You're welcome. Agree, it is :)

[@chamikanandasiri897]

Great Video. Very good explanation

[@memoai7276]

This is amazing and thank you for making these! Exactly what I was looking for. Very clear and precise! Many kudos :)

[@BotAcademyYT]

Memo Alci thank you! Glad that I can help you and others with my videos :)

[@ritikpratapsingh9128]

please keep on making these videos. i always wanted to learn this.

[@BotAcademyYT]

happy to hear! I'll definitely make more videos, but I might need a few weeks per video cause my master studies take quite a bit of my time till end of the year. Luckily my thesis is about Reinforcement Learning.

[@ritikpratapsingh9128]

@@BotAcademyYT that's great. Just keep uploading these fantastic videos. Thankyou

[@FabienChampigny]

Strait forward explanations, thanks !

[@jz9991z]

Sorry, just one more question. Is it possible to combine Imitation learning with Reinforcement learning? Like training an RL model with demonstration data?

[@BotAcademyYT]

yes, it is called behavioral cloning, make sure to check out the ML-Agents docs for more information about it

[@jz9991z]

@@BotAcademyYT thanks a lot

[@KarenKawas]

Amazing content. Subscribed!

[@krissloo143]

beautiful

[@nozercode]

9:59 ERROR: ml-agents.evns is not a valid editable requirement. It should either be a path to a local project or a VCS URL (beginning with bzr+http, bzr+https, bzr+ssh, bzr+sftp, bzr+ftp, bzr+lp, bzr+file, git+http, git+https, git+ssh, git+git, git+file, hg+file, hg+http, hg+https, hg+ssh, hg+static-http, svn+ssh, svn+http, svn+https, svn+svn, svn+file).

[@Derekstefan]

I know this was 4 weeks ago but in case you haven't been able to solve it. I'm assuming you are trying to do the -e ./mlagents-envs command. I fixed that by putting the path to where envs folder is.

[@anicsim8390]

its spelled envs not evns? try seeing if fixing your typo was the issue

[@jiantang6871]

Dose anyone get this problem where the Unity Console pop up and said :"Couldn't connect to trainer on port 5004 using API version 1.4.0. Will perform inference instead.",i google it but still couldnt fix it ,can anyone help me out here?ill b very appreciate it

[@jiantang6871]

i kinda solve it but dont no how,i reopen the project in unity,then use the ancaconda prompt ,and some how it work this time,im comfused but gald that it worked,anyway,it wont work if it wasnt your video showed me how to run it,great video!

[@jiantang6871]

dont no if its wat solve the problem，i realize that theres no port 5004 in my pc，so i created one for it ,but i tried to run it the moment i did it and it failed,so i thought i wasnt the solution and went 2 sleep，but this morning i woke up and test it again,it work prefectly,so i thought i could share this 2 anyone whos having the same issue as me

[@micabarshap]

I got error with --train , "Use --inference to run in inference mode." --train option has been deprecated. Train mode is now the default. Use --inference to run in inference mode." what to do help!!!

[@BotAcademyYT]

Oh interesting. Seems like you don’t need the train option anymore. Is it working without the -train?

[@electromorphous]

such cool, much thanks

[@biswasshubendu4]

can you make a detailed playlist? or can you refer some materials for learning these? i am beginner in unity and i know deep learning thank you

[@BotAcademyYT]

I'll definitely make more videos and create a playlist out of them. If you wanna learn Unity in general I'd recommend a Channel called Brackeys.

[@jz9991z]

Hi, just a quick question. If you stop the training process, can you resume it later?

[@BotAcademyYT]

Hey. Yes you can with a resume argument when re-running. I think I covered it indirectly in one of the later video (part 3-5)

[@jz9991z]

@@BotAcademyYT cheers

[@barberaTP]

Nice video! Btw, how do you manage to config your Gpu to train the model instead the cpu( default). Thanks

[@BotAcademyYT]

Thanks! I haven't set it up either (that's why I am getting the warnings too). The ML-Agents team wrote that using the GPU only makes sense when training on hundreds of environment copies with agents using visual observations. I'll make a video about how to set it up at some time in case someone really needs it.

[@barberaTP]

@@BotAcademyYT that would be nice. already subscribe your channel and looking forward for any GPU config updates! thanks

[@behos5719]

Cool, thank you for the tutorial!

[@leandro_damasceno]

Great content!

[@vedantbhagat5721]

thanks for teaching me how to install unity

[@BotAcademyYT]

You’re welcome. If you‘re already familiar with something, you can see the Chapters in the Timeline of the videos and jump to the chapter where you want to start. I‘ve done that in nearly every video so that people don‘t have to watch something that they already know. Hope you‘ll enjoy the other videos which will go deeper into ML-Agents.

[@Captinbubby4355]

ausgezeichnet!

[@ryaeon9793]

help, i dont have ml agent package on my unity list.

[@NewStoryline]

*Please, help!* When I run the mlagent-learn command, I get an error: ImportError: DLL load failed: ?? ?????? ????????? ??????. Failed to load the native TensorFlow runtime (Sorry for my English)

[@BotAcademyYT]

Are you using Windows 10 and python3.7 via an Anaconda environment? My first guess would be a wrong python version.

[@NewStoryline]

@@BotAcademyYT I use Windows 7 and python 3.7

[@BotAcademyYT]

I assume that it is not compatible with Windows 7 anymore. I don't know your situation, but I would highly recommend updating to Windows 10 (regardless of this video) because Windows 7 is not maintained by Microsoft anymore and might be vulnerable to security risks.

[@NewStoryline]

@@BotAcademyYT Thanks!

[@Draco98]

@@BotAcademyYT i'm also getting the same error and i have windows 10 and python 3.7

[@JUMPCUTART]

so cool!

[@ryaeon9793]

i done getting-started.md but im stuck at the last part.. where is "results" folder? to get that 3dballs.onnx files i just created ?

[@BotAcademyYT]

I think you're using a more modern version (and not Release 1). I think at some point the folder was renamed from results to something similar. If you would use Release 1 though, there should be a results folder as shown in the video

[@ryaeon9793]

@@BotAcademyYT yeah.. i use release 18.. im quite confused right now haha.. Try to ask discord and forum but no respond for now

[@BlackSheeeper]

seems like a good foundation for a tower defense game ;-)

[@ZoubaierAouadi]

Cant really see the use of it in a tower defense where towers can simple look and shoot at the assigned targets at runtime

[@foolarchetype]

hello, i have a problem mlagents_envs.exception.UnityTimeOutException: The Unity environment took too long to respond. Make sure that : The environment does not need user interaction to launch The Agents' Behavior Parameters > Behavior Type is set to "Default" The environment and the Python interface have compatible versions. Version information: ml-agents: 0.25.0, ml-agents-envs: 0.25.0, Communicator API: 1.5.0, PyTorch: 1.7.1+cu110 and second: Couldn't connect to trainer on port 5004 using API version 1.5.0. Will perform inference instead.

[@BotAcademyYT]

Hey, it seems that you've installed mlagents v1.5 in Unity. But you need v1.9. You can find the version match between mlagents python package and mlagents unity package under the 'Releases & Documentation' section in the github README (github.com/Unity-Technologies/ml-agents)

[@foolarchetype]

@@BotAcademyYTyour link is not work/ thx i will try

[@foolarchetype]

com.unity.ml-agents (C#) v1.9.0 i have the same/

[@BotAcademyYT]

@@foolarchetype github.com/Unity-Technologies/ml-agents

[@BotAcademyYT]

Hmm, the problem message is quite generic, so it's hard to figure out the problem. I'd suggest joining the Bot Academy discord server and asking there

[@iliketacoswithtacos]

after installing ml agent, and starting the project, it would not show up in my folders. I need some help with this, because it's really annoying and time consuming to figure out whats wrong

[@BotAcademyYT]

Hard to guess what’s wrong there. Could you try the steps exactly as described in the video? If it still doesn’t work I can offer you to help via Discord with screensharing

[@iliketacoswithtacos]

@@BotAcademyYT thanks, but I figured out whats wrong, the file wasn't unzipped so it didn't show up.

[@BotAcademyYT]

@@iliketacoswithtacos all right. Great that you figured it out :)

[@AdrienTriangles]

Is it possible to train them at runtime?

[@BotAcademyYT]

it is not possible at the moment. The algorithm is slightly different compared to training (no exploration, ...). Why do you want it to train at runtime? What you could do is resume the training of a trained model with --resume and then set the environment speed from 20x to 1x so that it is not sped up and looks like it's at runtime.

[@AdrienTriangles]

@@BotAcademyYT Oh I get you ! I was thinking it would be lit to let players train them or have them to train against players ^^ Still great !

[@BotAcademyYT]

ah I see. It could actually work with environment speed set to 1 and some effort to set it up. Haven’t tried though. You could also take a look into behavioral cloning where AI learns based on your own recordings :)

[@alvaromachucabrena97]

Hello ! Do you know what type of activation function use ml-agents? And how configurate it?

[@BotAcademyYT]

Hey. Good question. From what I know, they're using the recommended activation function for each Algorithm / configuration. You can just change it by changing their implementation (there is no way to change it through the config). They've a file called tensorflow_to_barracuda where I found the following: known_activations = { "Linear": 0, "Relu": 1, "Softmax": 2, "Tanh": 3, "Sigmoid": 4 } This might be a list of all activation functions that are somewhere used in their implementations.

[@alvaromachucabrena97]

@@BotAcademyYT Thanks a lot, bro !!

[@rohitk6817]

It s an amazing well explained video, this is what I was looking for . I am getting an error though after I typed mlagents-learn. Error:- importerror: dll load failed: the specified module could not be found. failed to load the native tensorflow runtime. Could you help me please.

[@BotAcademyYT]

Thank you! Could you try installing Visual Studio 2019 from visualstudio.microsoft.com/downloads It's anyway needed for the second video and will install required dependencies that tensorflow needs (C++ Redistributables). Good chances that it'll resolve the issue. If you just want to install the required C++ Redistributables, you can go to this webside: support.microsoft.com/en-us/help/2977003/the-latest-supported-visual-c-downloads You need to download either the x86 (32 bit) or x64 (64 bit) version depending on your system. If you don't know whether you have 32 or 64 bit you can open a command promt and type 'wmic OS get OSArchitecture'. If that doesn't work, downgrading Tensorflow from Version 2.1 or 2.2 to 2.0 (via pip) worked for another viewer.

[@tanzimulittahad623]

Showing this error please help me ImportError: DLL load failed while importing _pywrap_tensorflow_internal: A dynamic link library (DLL) initialization routine failed. Failed to load the native TensorFlow runtime.

[@tanzimulittahad623]

will you not help me ???? please help me

[@BotAcademyYT]

This should be resolved by downgrading Tensorflow with either pip install tensorflow~=2.1 or pip install tensorflow~=2.2

[@tanzimulittahad623]

@@BotAcademyYT still showing the same problem ImportError: DLL load failed while importing _pywrap_tensorflow_internal: A dynamic link library (DLL) initialization routine failed. Failed to load the native TensorFlow runtime. please help me directly by taking my pc remote control by teamviewer . I will remain very grateful to you .

[@BotAcademyYT]

Please make sure to check all comments and probably join the discord server. The issue can have multiple causes. Most of them are answered in the comments here or on discord. Chances are good that you‘ll find a solution there. If not, please ask in the discord server. If no one there knows a way to solve it I can take a look

[@aryan_kode]

hey they have a release 2 should we download that?

[@BotAcademyYT]

Release 2 includes just some minor changes to Release 1, so it should work. Please let me know if it somehow doesn't.

[@sohanmahmud5124]

hi. after completing your study which direction you want to go? like unity development, web developer?

[@BotAcademyYT]

Hi. I want to focus on Automation Software & A.I. by combining them. A.I. focus will be on mostly on Reinforcement Learning (also writing my thesis about that - starting in a few month).

[@eugenemettrix4756]

Hi, when I want to train the AI, I get an error: « zsh: command not found: ml-agents », but when I type « conda list » I have the same output as you so I don’t understand where the error is from? I have the mlagents release 4 maybe it might be one of the reasons?

[@BotAcademyYT]

Make sure that the command starts with 'mlagents-learn' and not 'ml-agents'. Release 4 should be okay if you use the updated path to the config file (Example for the 3DBall environment with Release 4: 'mlagents-learn config/ppo/3DBall.yaml --run-id Whatever'

[@bugz9926]

a tutorial for mac?

[@BotAcademyYT]

Should basically be the same I think. If you have any questions or encounter any problems let me know. I have a Mac and can test it then.

[@mahmoudhazem1785]

can you create your own environment and your own neural network code or is it predefined neural network and you can't modify it ? @Bot Academy

[@BotAcademyYT]

Own environment yes definitely. Regarding the neural network code: The ML-Agents Team implemented the PPO and SAC algorithms in python with Tensorflow and linked them to the unity logic with a connector. So you can either use them with a bunch of config options (i.e. number hidden units, ...) or create your own algorithm / neural network & link it to the connector. This document explains it in more detail: github.com/Unity-Technologies/ml-agents/blob/master/docs/ML-Agents-Overview.md

[@mahmoudhazem1785]

@@BotAcademyYT thanks very much ,keep up the great work

[@joshuasteward6672]

but...how do you tell it what to train for? Seems to me like that would be the most important part.

[@BotAcademyYT]

This is explained in the second video. Make sure to check it out :)

[@lukenukem8028]

This is all well and good in the context of current projects which have goals, like keeping the ball on the cube. Would be good to see how you set up all the project goals first, otherwise, how does the 3rd party app know what it good and bad.

[@BotAcademyYT]

Thanks for you comment. I created a second video where I set up a project from scratch and went through all steps like defining the reward. I am now sure what u meant with 3rd party app. If you mean Unity and the Python library, they are connected through a python connector defined in the python library.

[@lukenukem8028]

3rd Party: Yeah that whole Python thing is not a part of the Unity parkage, so it's called 3rd party.

[@lukenukem8028]

So it's not the reward that matters so, much since that again replies on a project. What I mean is, just Unity etc. The Cube scripts that tell it to tilt in every direction to keep the ball on it. The game engine still has to be built first before bothering with creating Python memories etc.

[@jibinjohn8263]

If I am creating a python ML model how can I use it inside Unity? Example A face recognition model. I don't know is this the exact thread to ask this question.

[@BotAcademyYT]

First, you can not directly use that model to interact with the environment (only .nn models trained through unity can directly interact with the environment). But I guess that is not what you're planning to do. If you want to use the model in addition to the reinforcement learner to improve its observations by adding additional information through a face recognition model or similar, that should work (and makes sense). From what I know, you have to use the ml agents python connector and (I guess) side channels. Haven't used them so far. In any case, you need to go a bit into ml agents code, so I would recommend going through the ml agents documentation and code. github.com/Unity-Technologies/ml-agents I hope I could somehow guide you in the right direction.

[@jibinjohn8263]

@@BotAcademyYT Thanks I will give a try

[@danielegiunchi9741]

if I want to create a new project, the name in the config file needs to be equal to the Unity scene?

[@BotAcademyYT]

It needs to be equal to the name in the Prefab. For example in the Soccer example is the Prefab "StrikersVsGoalieField.prefab". If you click in there you'll see the GameObjects like "BlueStriker" or "PurpleGoalie". When you click on them you can see in the Inspector that they have a Bahavior Name. That is the name you need to set in the config.

[@danielegiunchi9741]

@@BotAcademyYT nice thanks! could you do a video where you explain the way to put a custom Tensorflow model instead of the one created by mlagents?

[@BotAcademyYT]

You're welcome :) Directly using a TensorFlow model that was trained somewhere shouldn't work. It has to be trained with Unity & ml-agents. If you mean creating an custom algorithm with TensorFlow and use this custom algorithm with ml-agents to train the model in Unity, yes - good idea! I might need to do this for my thesis in a few month anyway, so I could make a video then. Might take a few month, though. First, I'll create a few videos to go through the basics.

[@tanzimulittahad623]

Showing these error pkg_resources.ContextualVersionConflict: (numpy 1.19.1 (e:\anaconda\envs\unity\lib\site-packages), Requirement.parse('numpy=1.16.0'), {'tensorflow'}) please help me............................ please. please.

[@BotAcademyYT]

You need to downgrade numpy. I think the command was ‘pip install numpy==1.18’ with either 1 or 2 ‘=‘ signs. Does that work for you?

[@tanzimulittahad623]

@@BotAcademyYT thanks I will try

[@tanzimulittahad623]

@@BotAcademyYT Thanks . I will try..

[@tanzimulittahad623]

what does it mean ???? (unity) PS C:\Users\Ittahad's pc> pip install numpy==1.18 Collecting numpy==1.18 Downloading numpy-1.18.0-cp38-cp38-win_amd64.whl (12.8 MB) |████████████████████████████████| 12.8 MB 65 kB/s Installing collected packages: numpy Attempting uninstall: numpy Found existing installation: numpy 1.19.1 Uninstalling numpy-1.19.1: Successfully uninstalled numpy-1.19.1 Successfully installed numpy-1.18.5 (unity) PS C:\Users\Ittahad's pc> What does it mean ?.........

[@tanzimulittahad623]

What does it mean tensorflow 2.3.0 has requirement numpy=1.16.0, but you'll have numpy 1.19.1 which is incompatible.

[@praveshgunnuck2922]

what if I have python 3.8? Does that make any difference ? '

[@BotAcademyYT]

It shouldn’t make any difference. Let me know if you run into an error. Would be interesting to know.

[@praveshgunnuck2922]

@@BotAcademyYT Thanks that worked great.. Earlier I was trying to install mlagents without the repository that is typing pip install mlagents And I don't know why there was something timed out during the process.. But now that I tried installing from the repository, it worked totally fine. I can finally complete my Final Year Project. YOSHAAA!!!! Thanks for the tutorial.

[@BotAcademyYT]

Awesome! Thanks for letting me know. :) All the best for your project! You definitely got an interesting one if it’s in that area!

[@bird.7688]

Omg Mannheim?! Bin voll in deiner Nähe! Bin 16 Jahre alt und will in die richtung Arbeiten gehen, kannst du mir sagen was ich arbeiten will? Ich will halt Games Designen oder Coden :( Hast du discord?

[@BotAcademyYT]

Hey Niko. Cool! Würde erstmal mit Coden anfangen, das schadet in keinem Fall. Habe nen Discord Server: discord.gg/6fRE4DE (dort aber bitte auf englisch schreiben). Privat kannst mir natürlich auf deutsch schreiben.

[@geckoboi2029]

where is the anaconda promt on mac?

[@geckoboi2029]

Bot Academy pls help

[@BotAcademyYT]

you don't need it on mac. Just use terminal

[@geckoboi2029]

@@BotAcademyYT oh ok thanks :)

[@geckoboi2029]

@@BotAcademyYT also, is this the same for all other examples or just the 3d ball

[@BotAcademyYT]

You can run all examples with this setup :)

[@lesmilansdevany8014]

I don't have any trainer_config.yaml file in my config folder :( (i have the ml-agents-release_3) So i created it manually, but when i run the command : "mlagents-learn config\trainer_config.yaml --run-id=first_run-3dball --train" it gave me this error message "mlagents.trainers.exception.TrainerConfigError: The option 3DBall was specified in your YAML file, but is invalid." So i used the 3DBall.yaml file in the folder "config/ppo" and it worked ^^ (mlagents-learn config\ppo\3DBall.yaml --run-id=first_run-3dball --train)

[@BotAcademyYT]

That's because they changed the structure of the configuration with Release 3 (I'll make a video about that soon). It'll work with Release 1 and 2. If you want to use Release 3 you need to use another path to the trainer config file (just look in the config folder to find the correct one). The second issue is easy to resolve. It is not "ml-agents-learn", you need to write "mlagents-learn".

[@lesmilansdevany8014]

@@BotAcademyYT it's ok, i fixed all the problems and edited my message. Thank you for all. I the way you explain.

[@etto4425]

Hello sir! I love your series about AI in Unity! The only problem is that it gives me an error when I type mlagents-learn config\trainer_config.yaml --run-id=first_run-3dball --train, it gives me a very long error, but I got the most important part: FileNotFoundError: [Errno 2] No such file or directory: 'configtrainer_config.yaml' Does someone know a fix? Thanks a lot! P.S I'm on Mac

[@BotAcademyYT]

Hey and thanks! If you are using mlagents release 3 or higher, you need to adjust the path to something like config/ppo/... (just check the folder structure). If you're using release 1 as explained in the video you might just change the backward slash to a forward slash (probably its interpreted as escape character - not sure though). In any case, just check the folder structure and check if the file is there

[@etto4425]

@@BotAcademyYT Wow man! Thanks a lot it worked! But it must be my ability at coding, because I tried replicating the chrome dinosaur game, I added a raycast 2d etc... but my AI isn't just learning! Do you have any solution?

[@etto4425]

He just keeps jumping and never improves!

[@BotAcademyYT]

I've no idea to be honest. You might wanna join my discord server with around 70 people. Probably someone else might be able to help you there. Its also easier to post photos etc.

[@hoangminh6899]

great, make some Eco-System bro

[@nalydsworkshop]

It just gave me 153 errors. How do I solve them

[@BotAcademyYT]

Can you send me a screenshot with the error to [e-mail]?

[@danthiel8623]

BOTS be looking kinda THICC

[@NUL-SIL]

dani

[@ishkapoor7931]

Please apply some ML/AI in Godot Game Engine as your new tutorial series. Would be a great help!!!

[@BotAcademyYT]

Thanks! I added it to my list. But it might take some time. Schedule for this year is quite full (mostly with KZbin unrelated tasks). That will change next year though.

[@PetroZilia]

Great video! can you please make one when you so how to create and train an environment from scratch?

[@BotAcademyYT]

Thanks! Already working on that :)

[@krissloo143]

We'd appreciate a full series on making our own projects from scratch with python

[@BotAcademyYT]

Thanks for letting me know. I have plans on doing that, but a Unity environment has to be created in C# or JavaScript. Python can only be used for a custom Machine Learning algorithm. Will also make a video about that, but that might take a few month cause I wanna first cover the basics.

[@krissloo143]

@@BotAcademyYT Yes, I know, by python I meant the machine learning process :3, Imma follow the basics along and wait for these months to pass, best luck

[@martasandnes4251]

it didnt work it said it didnt work

[@BotAcademyYT]

If you let me know what error you got I might help you

[@Emperors_A7MET]

Redoing it again after 2 year's ,😂

[@tanzimulittahad623]

when i command (unity1) C:\Users\Ittahad's pc>pip install mlagents the result came like this ERROR: After October 2020 you may experience errors when installing or updating packages. This is because pip will change the way that it resolves dependency conflicts. We recommend you use --use-feature=2020-resolver to test your packages with the new resolver before it becomes the default. tensorflow 2.3.1 requires numpy=1.16.0, but you'll have numpy 1.19.2 which is incompatible. mlagents-envs 0.20.0 requires numpy=1.14.1, but you'll have numpy 1.19.2 which is incompatible. Successfully installed Pillow-7.2.0 absl-py-0.10.0 aiohttp-3.6.2 astunparse-1.6.3 async-timeout-3.0.1 attrs-20.2.0 cachetools-4.1.1 cattrs-1.0.0 chardet-3.0.4 cloudpickle-1.6.0 gast-0.3.3 google-auth-1.22.0 google-auth-oauthlib-0.4.1 google-pasta-0.2.0 grpcio-1.32.0 h5py-2.10.0 idna-2.10 keras-preprocessing-1.1.2 markdown-3.2.2 mlagents-0.20.0 mlagents-envs-0.20.0 multidict-4.7.6 numpy-1.19.2 oauthlib-3.1.0 opt-einsum-3.3.0 protobuf-3.13.0 pyasn1-0.4.8 pyasn1-modules-0.2.8 pypiwin32-223 pywin32-228 pyyaml-5.3.1 requests-2.24.0 requests-oauthlib-1.3.0 rsa-4.6 six-1.15.0 tensorboard-2.3.0 tensorboard-plugin-wit-1.7.0 tensorflow-2.3.1 tensorflow-estimator-2.3.0 termcolor-1.1.0 urllib3-1.25.10 werkzeug-1.0.1 wrapt-1.12.1 yarl-1.6.0 (unity1) C:\Users\Ittahad's pc> please help me ...what to do

[@BotAcademyYT]

Just run pip install numpy~=1.18 before installing mlagents.

[@laserflip]

yo tysm i was searching up A.I. training I found nothing exepect urs

[@timbeier2337]

Ich wusste doch ich kenne den Akzent xD

[@BotAcademyYT]

da hab ich mir soviel Mühe gegeben, dass man es nicht raushört - wird wohl noch ein Weilchen dauern :D

[@danthiel8623]

Be looking kinda THICC

[@Emperors_A7MET]

i am having problems with anaconda

[@BotAcademyYT]

hey. I can try to help if you join the discord server and explain your problem with a few screenshots

[@Emperors_A7MET]

Ok

[@Emperors_A7MET]

I have joined your discord and send a screen shot in the ml agents page in your discord pls reply

[@tarkozyol1284]

Bro uses cheat engine

[@McPerest]

I am 20 and I am still in 1st year Bachelor lol