Improve your Unity A.I. | Hyperparameters

Подписаться 11 тыс.

Просмотров 4 тыс.

50% 1

This video is about the Hyperparameters that can be set for the Reinforcement Learning algorithms Proximal Policy Optimization (PPO) and Soft Actor-Critic (SAC), which are two commonly used state-of-the-art Reinforcement Learning algorithms.
Please let me know if you have any questions. I'll try my best to help you out.
PPO Paper:
arxiv.org/abs/1707.06347
SAC Paper:
arxiv.org/abs/1801.01290
Find me on:
Discord: / discord
Twitter: / bot_academy
Instagram: / therealbotacademy
Patreon: / botacademy
Download my Unity AI example:
github.com/Bot-Academy/BallJump
Credits:
15.54 - End
▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬
Music: Ansia Orchestra - Hack The Planet
Link: • Ansia Orchestra - Hack...
Music provided by: MFY - No Copyright
▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬
Contact: smarter.code.yt@gmail.com
Chapters:
00:00 Intro
00:17 Config: Release 1 vs. Release 3
01:00 Parameter: learning_rate
02:02 Parameter: learning_rate_schedule
02:36 Parameter: batch_size
03:55 Parameter: buffer_size
05:00 Parameter: beta
06:48 Parameter: epsilon
08:05 Parameter: lambda
09:20 Parameter: num_epoch
10:31 Parameter: buffer_init_steps
11:36 Parameter: init_entcoef
12:16 Parameter: save_replay_buffer
14:18 Parameter: steps_per_update
15:19 Parameter: tau
15:44 Outro

Опубликовано:

26 июл 2024

Ссылка:

Скачать:

Готовим ссылку...

Добавить в:

Мой плейлист

Посмотреть позже

Комментарии : 17

@ioda006 3 года назад

So few videos out that explain all the parts of ml-agents. Thank you so much for putting these together, this really helps.

@rajeevjawaji4472 4 года назад

this is really good, i would recommend this to anyone strugelling

@asa989asa 3 года назад

Great video. Thanks really. I think many ppl would appreciate if you can make a video about other parameters like network settings, reward signals and specially GAIL and BC. Your graphs are extremely valuable for us to compare the results for each different settings. Looking forward for more ML videos! :)

@pixies64 3 года назад

wish i could give you more than one like

@BlackSheeeper 4 года назад

really helpful

@bnok5237 4 года назад

THANK YOU FOR THE LESSON ALOT , but i have some questions on how to get the graphs , where can i find them and how can i see them

@BotAcademyYT 4 года назад

Glad you like it :) Check out my tensorboard video. In there I explain everything related to the graphs

@alvaromachucabrena97 3 года назад

Hello, How can I know which is the best neural network architecture to use? for example if I used two different models.

@BotAcademyYT 3 года назад

Mostly through experimentation and looking how the training progresses/how quick the agent learns

@aryan_kode 4 года назад

do a video on coding algorithms to train unity environments

@BotAcademyYT 4 года назад

good idea. Will do that at some point. But I'll probably make a few Python videos first.

@alvaromachucabrena97 3 года назад

Do u know if can I use a method like GridSearch in python to do the tuning for my model?

@BotAcademyYT 3 года назад

From what I know it should be possible, but not so easy. But I think unity has something like that. I saw a video where someone let run horses through a maze using a similar algorithm. Just try to search for “unity maze horses” or so on RU-vid. The channel name was brackeys.

@alvaromachucabrena97 3 года назад

@@BotAcademyYT Thanks a lot !!

@GaithTalahmeh 3 года назад

Thank you alot Sir, but can you please make videos about AI in general, Like Building neural networks and these things Apart from that, Great video , Thank you for investing in the ad, you earned a sub!

@BotAcademyYT 3 года назад

Thanks! Planned for future videos :)

@GaithTalahmeh 3 года назад

@@BotAcademyYT Thanks!