make epsilon greedy sampler support GPU mode

facebookresearch / ReAgent

A platform for Reasoning systems (Reinforcement Learning, Contextual Bandits, etc.)

https://reagent.ai

BSD 3-Clause "New" or "Revised" License

3.58k stars 521 forks source link

Closed alexnikulkov closed 2 years ago

alexnikulkov commented 2 years ago

Summary: In the GPU mode, the function will raise error since tensors created by torch.ones or torch.zeros are in CPU and rest tensors are in GPU.

Differential Revision: D41062175

facebook-github-bot commented 2 years ago

This pull request was exported from Phabricator. Differential Revision: D41062175