Fix RNNs with DDP - Githubissues

PG642 / multi-sample-factory

High throughput reinforcement learning on clusters

MIT License

2 stars 0 forks source link

Closed KonstantinRamthun closed 2 years ago

KonstantinRamthun commented 3 years ago

Currently we can't use RNNs with DDP. This should be fixed to restore the original functionalities of sample-factory.