This project is all about training a Deep Q-network to play flappy bird and hopefully be good at it.
used claude code for building the interface
I believe ideas don’t mean anything until you can experience/feel them in some way or the other.
That’s why i built an interactive live web demo of my trained model playing flappybird in real time.
If you liked the demo, do share your feedback! <3.
PS : Demo link - https://huggingface.co/spaces/zanwok/flappy-bird-dqn
Log in to leave a comment
Tried to build an interactable exe where i could see the ai playing live… the last try resulted in a 2.8 gb exe file. lol
Is this “wrong”?
No.
This is exactly what happens when you try to ship PyTorch + a RL environment as a standalone exe.(which doesnt work)
Log in to leave a comment
Trained the Deep Q-Network as best as i could in a day to play flappy bird
Log in to leave a comment