Running on a GPU
chokosabe opened this issue · comments
It isn't clear whether this can run using a gpu. Tried with the tf-nightly-gpu build and I'm getting an error.
It isn't clear whether its the library or tensorflow itself. The environment is colab.
The error itself is detailed here:
Seems to be somewhere between TF and TF probability. EasyAgents is sitting on top of TF-agents, which in turn sits on TF. So I guess our chances are small to help from the EasyAgets side. Or do you have an suggestions ? In any case, we will try to update the direct dependencies of EasyAgents the next few weeks. So maybe that might then also have a "beneficial" impact.
I know for sure that using tf agents installed by easyagents as a
dependency runs on GPU if there is one. Wouldn't be able to think of any
reason why easyagents itself should interfere.
…On Sun, 6 Sep 2020, 20:18 Christian Hidber ***@***.***> wrote:
Seems to be somewhere between TF and TF probability. EasyAgents is sitting
on top of TF-agents, which in turn sits on TF. So I guess our chances are
small to help from the EasyAgets side. Or do you have an suggestions ? In
any case, we will try to update the direct dependencies of EasyAgents the
next few weeks. So maybe that might then also have a "beneficial" impact.
—
You are receiving this because you are subscribed to this thread.
Reply to this email directly, view it on GitHub
<#111 (comment)>,
or unsubscribe
<https://github.com/notifications/unsubscribe-auth/AADAJKL2NSP42QQLQ4UMTXDSEPHBBANCNFSM4QUZSC7A>
.
Yeah this is looking more and more like a Tensorflow issue. Tried in Rllib and getting the same thing.