How to train your bot’s micro from scratch
|
Training micro with reinforcement learning sounds like something only Alphastar could do. But what if you could start with just one unit and two actions? SharkNice showed how to easily train unit micro from scratch using Sharky RL Matrix Training (or Stable Baseline 3 for python) , starting with a single Stalker kiting two Roaches and scaling up from there. But the key take away though : scope your action space to preset functions your bot already knows. Instead of giving RL infinite options (“move anywhere on the map”), you give it a handful of choices it can actually learn from: attack closest, retreat, flank. The AI picks between known-good behaviors rather than reinventing movement from scratch. This came out of our Postbots session on Training Reinforcement Learning Micro From Scratch With Sharky RL Matrix. Look at your bot this week — find one unit behavior you already have coded as a function, and think about letting RL choose when to use it. [Come join us for the next one ->]
|