Video

Thanks for the Nice Video

How is it useful?

The second chapter's title should be "Cart Pole", not "Cart Pool"!

This is one of the best tutorial for DQN. Thanks a lot man

this guys talks way to fast

When I run your notebook on my machine the training time is so much slower! You are able to do 1000 episodes in <30 seconds in the video but on my own machine it takes more like 45 minutes! Do you know what could cause this massive increase in runtime?

In this video Jon mentions a previous video where neural networks are already introduced. Where is that video? Cannot find it in this playlist since current video seems to be the first one in this playlist?

how to download the slides?

How do you select the pivot of rotation

1:04:10 We are still performing gradient descent, rather than gradient ascent, because we're using the Mean Squared Error loss function, between a target q-value prediction, and our current q-value prediction. Therefore the further away we are from this target, the larger the Mean Squared Error, and that distance is what the stochastic gradient descent is minimising.

After watching this video it seems like I have been taught RL by Johnny Sinns😂

Thank you for these great lessons!

Thankyou , I almost gave up trying to understand this. Big thanks

Loving your Mathematical Foundation series on Udemy right now. Planning to hit up O'Reilly as soon as I finish these so I can continue with your curriculum.

Thank you for this excellent tutorial! Better than several others I watched and didn't find as helpful. The way you built things up theoretically and in code was extremely well organized, well explained, and easy to follow. Much appreciated!

Day saver! Thanks a lot.

Very nice explanation, thank you very much

Best explaination, after seeing so many channels videos 🙏

I read it. It has REALLY good advice. Half of it is just filler though; Uncle Bob's flavour of Agile is simple enough to be explained in 50-60 pages, so there's a bunch of extra unnecessary stuff there, even for his usual standards, to hit the 150+ pages that warrant it being a full-priced book. But the 50-60 pages that actually get into the meat of it are REALLY good, don't get me wrong.

I tried to run the code but failed with "ValueError: setting an array element with a sequence. The requested array has an inhomogeneous shape after 1 dimensions. The detected shape was (2,) + inhomogeneous part." on "state = np.reshape(state, [1, state_size])", may I know why? Thank you very much!

Amazing explanation, thanks!

This is too good!!!

iOKPowerPoint is not a word. It is a name. Being idle too long.

Thanks that was helpful

Hello Jon, I am not finding the code in your Github.. can you please help me. I will be thankful

Simple explanation..you helped me few hours of reading tough paper.thanks

Excellent tutorial. I am struggling with the explanation on the target_f and the mapping. Can someone clarify the intuition of that step? It is clear what it does, but I don't get why is this useful.

The best teacher

best java tutorial

I had a mini city app in the side bar. I really loved it being there, but some kind of other download that helped power it went defunct, so now it has gone.

Amazing tutorial. I couldn't imagine of learning complete DQN tutorial in an hour with so precise code.

Hi, first of all, thank you for this great video!!!! Unfortunately i always get the same error message when testing: ValueError: cannot reshape array of size 2 into shape (1,4) I can't get this solved, can anyone help? Thanks in advance!

Good stuff!

52:24 - you wrote something extremely confusing in probably the most important part of this tutorial! why would you calculate target_f=self.model.predict(state) and then replace the most important value of it with target in "target_f[0][action]=target?!? I'm really banging my head on the table here

Hey sir, very good tutorials. Pls i'm having issues with docker, i make couple of research and found out that i need to enable virtualization on my PC; I try enabling it ,even went as far as contacting the laptop brand support but no assistance was working. Pls, is there any alternative that you can suggest for me.. This whole stuff is becoming frustrating🙏

Really great explanation! Thank you so much @Jon

I just need one clarification regarding this. How much time did it take to reach 1000 episodes ? It seems to be fairly quick. I am trying to solve it but it takes me around two hours. I don't know why. Maybe, the dependencies, the versions are not compatible with each other but I am not sure. Can you please help and clarify this for me ? I am using tensorflow 2.9, ide is pycharm and my processor is Intel i7 11800h. I've tried it using gpu as well but it takes me 1-1.5 hours.

Muhteşem anlatım - Excellent presentation !

Thanks for the lecture. Very educative. I am confused by the gradient ascent. How does the keras fit know it is supposed to do gradient ascent and not gradient descent?

In CartPole-v1 the maximum length of an episode is 500 so the bar has been raised now.

Iam looking on this video 2022

Mister Job aye

seriously this is one of the best videos which i have seen in youtube, thank you so much!

I am looking for the other two videos, anyone can point me to them???

Can you explain how you transform high-level ideas outlined in a product roadmap into definitive items in a product backlog?

Would it be possible to star the software used for that simulation. I am trying to do something similar using keysight Genesys and EM pro with very limited time to learn the ins and outs of those softwares. Thank you.

New Word is shit ... old Word is great.

Hi, Thank you so much for the great video, I can see from your talk is the agile coach at the team level is actually the scrum master, should be the team level coach be the same as the scrum master or a different role called Team Agile Coach

That is exactly what I have been looking for! Thanks a lot!

Nice