Gianni De Fabritiis

New feature: Save trajectories incrementally

1

At the moment we use np.save(of the entire trajectory), we need to be able to append safely (I mean safely because I don't trust h5py to be able to do...

PyTorchRL: Modular and scalable reinforcement learning in pytorch reaching new high in obstacle tower challenge

3

Hi, don't know if it is the right place, but it is probably relevant for anyone interested in obstacle-tower-env. https://arxiv.org/abs/2007.02622 We have scaled up our second position approach in the...

@ikostrikov We have developped a new library starting from this one. We have submitted now what we think that is a practically final version. It is still based on pytorch...

I have talked to Albert to help