search

jurgisp / memory-maze

Evaluating long-term memory of reinforcement learning algorithms
MIT License
128 stars 13 forks source link

Data generation #29

Open Frederik-Alexander opened 8 months ago

Frederik-Alexander commented 8 months ago

I was wondering if you could post the code of how you generated the data, as I'd like to do the same with some modifications. With my naive data generation approach it seems the agent looks at walls most of the time.

Regards

gunshi commented 4 months ago

+1