Sasha Abramowitz
Sasha Abramowitz
Hey @Howuhh really sorry for only getting back to you literally a year later. Really great work on xland minigrid, it's an awesome benchmark! I'd be open to this although...
Easier to review the diff between masac and hasac: https://www.diffchecker.com/EKMo2FcC/
Some additional testing shows that tfp and numpyro produce the same out puts, needs more investigation to 100% confirms this, eg for larger sample shapes and through backprop ```python import...
> So basically, because the actors are faster now, the pipeline fills up faster and then causes the insertions to slow down a lot i think because there is more...
Honestly I'm not convinced we even need the shard map. You're welcome to close this PR if you want
Looks like you're not donating the correct argument, argument 0 is self, you need to donate the buffer state which is argument 1. You can also use `donate_argnames` to be...
Ah my bad, I was on my phone and I thought you were jitting the class method. I had a look and it seems to be the fact that it's...
Just for the record this is the speed I get with a `TrajectorBuffer`. However I needed to switch to the same config as in the docs, when I use your...
Thanks for all the info @EdanToledo. I don't think I'll have time for the next month to look into this, but I can try in September. Possible that @SimonDuToit may...
Totally agreed @mickvangelderen. Would you be able to make a PR for this? :smile: I think here: ``` @pytest.fixture() def max_length() -> int: return 256 @pytest.fixture() def fake_transition() -> FbxTransition:...