Managing number of steps in a batch

DeepX-inc / machina

Control section: Deep Reinforcement Learning framework

MIT License

279 stars 45 forks source link

Managing number of steps in a batch #119

Open rarilurelo opened 5 years ago

rarilurelo commented 5 years ago

iterate_rnn in Traj class makes iterator of batches. A tail of the batches are zero padded for arranging length of episodes. For this reason we couldn't control number of steps in a batch.

rarilurelo commented 5 years ago

This problem can be solved by incorporating random_batch_rnn to iterate_rnn.