Matthew Carbone

@saitcakmak funny that I come back to this exactly a year later. Have there been any updates on merging #1372? 😁 Looks like the GPyTorch side of the changes is...

@esantorella ok no problem. I'm happy to help contribute somehow if there's any opportunity. Thanks! 👍

@takafusui incidentally, this > I am also reluctant to set a large beta in the UpperConfidenceBound analytic acquisition function Does actually work just fine. I usually use beta=100 or so...

@takafusui so in the case of just setting beta to be large, how large depends on your data. You need beta to be large enough to "outcompete" the value from...

@eytan no, but I've benchmarked it against my own custom acquisition function which does pure active learning and they work about the same.

Awesome, thank you, Dan! 👍

@danielballan Unless I'm mistaken this appears to now be fixed! (independent of #296?) ```python CLIENT = from_uri("https://aimm.lbl.gov/api") CLIENT.login() # ... # RuntimeError: API key authentication is being used. ``` And...

@ashleve It's really amazing of you to provide this as is and without a license, but yes based on my experience/understanding, your code is licensed under MIT and as such...

@ashleve not related to this issue, but I just wanted to say thank you for such a wonderful template. I have turned this into a [project](https://github.com/matthewcarbone/Crescendo) for training ML models,...

@ashleve it's strange because this output is from Lightning, not Hydra. And the Lightning output seems to be piped to neither `stdout` nor `stderr` (I've tried catching both). Anyways, I'll...