Impala CNN model - better performance for Reinforcement Learning - Part 1
00:00 Start
05:00 Pong - Atari 2600
07:21 Bezerk - Atari 2600
09:45 Mortal Kombat II - Genesis
13:03 Super Mario Bros. 2 Japan - FC
15:51 Memory Usage of Impala CNN
OpenAI CoinRun:
https://openai.com/blog/quantifying-g...
DeepMind's Impala:
https://arxiv.org/abs/1802.01561
Vanishing Gradient problem:
https://towardsdatascience.com/the-va...
My blog:
https://www.videogames.ai/2021/08/21/...
Add my game to your wish list on Steam!
https://store.steampowered.com/app/10...