notwa/smbot
1
0
Fork 0
Code Issues Pull requests Releases Wiki Activity

update todo

Browse source
This commit is contained in:
Connor Olding 2018-03-23 09:33:02 +01:00
parent 57cead431d
commit 4765104c7a
1 changed files with 3 additions and 1 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

4
_NOTICE
View file

@ -5,12 +5,14 @@ however, feel free to copy any snippets of code you find useful.
TODOs: (that i can remember right now)
- finish implementing backprop
- replace evolution strategy algorithm with
something that utilizes gradients like PPO
something that utilizes backprop like PPO
- settle on a network architecture
- normalize and/or embed sprite inputs
- fix lag-frames skipped-inputs bug
- detect frames when Mario is in a controllable state
- fix offscreen sprites sometimes being visible to network
- add some detection for enemies later in the game
- compute how many input neurons the network needs instead of hardcoding
naive:
- learn any combination of buttons, starting from title screen