Sam Foreman

> > Got stuck when compiling the fused_kernels when training on multiple nodes. But it works well in a single node. Why? > > Same here. Did you solved this...

awesome to hear, will try this, thanks!

@awan-10 @jeffra would it be possible to get a review on this (should pass + be straightforward)

If you're interested, I have a (slightly modified for personal use) upstream version in my personal fork that I've been working on that _should_ look okay on both light and...

Hello, I've used the gist you submitted to test the GMM model, and I believe the issue with the gist you provided is on line 81, where you have: `samples...

Ahh you're right actually, I apologize. I'll have to double check (I didn't save the times from my last run) but I believe using 10,000 training steps with a batch...

@Razcle I am running on CPU. I used temperature annealing, using @daniellevy 's suggestion starting at T = 5 and decreasing by 0.98 every 100 iterations. For the intermediate chain,...

@jxy is correct, using a higher starting temperature, `T = 10` in this case solves the issue and even after just 5000 training steps the trajectories are able to tunnel...

@jxy @Razcle Can confirm, after using an initial temperature of `temp = 40`, and training for 10,000 training steps I've found that the learned trajectories are able to tunnel between...

The variance is still 0.1, I've pushed the notebook for running the experiment to my forked repo, you can get the jupyter notebook for the MoG model [here](https://github.com/saforem2/l2hmc/blob/master/MOGExperiment.ipynb). I also...