Implementation of RL2 algorithm with PyTorch

Hi guys, I just implemented the RL2 algorithm (https://arxiv.org/abs/1611.02779) with PyTorch. The code is here: https://github.com/fatcatZF/RL2-Torch . I used a shared GRU feature extractor, with separate MLP heads for actor and critic. The neural network was optimized with the PPO algorithm.