Greener Deep Reinforcement Learning: Analysis of Energy and Carbon Efficiency Across Atari Benchmarks
Greener Deep Reinforcement Learning: Analysis of Energy and Carbon Efficiency Across Atari Benchmarks
arxiv.org
Greener Deep Reinforcement Learning: Analysis of Energy and Carbon Efficiency Across Atari Benchmarks
RL is still nowhere near the scale where this matters, but it's always good to get an idea of how things are going to look when they inevitably reach that point.
One thing that stood out to me was that none of their agents performed well on boxing or pong, but in my experience, these are the two easiest tasks in the entire Atari suite. They're the games I use as quick sanity checks. How did they manage that?