Ben Glickenhaus

icon location Ben Glickenhaus is a graduate student at UMass Amherst focusing on AI and Machine Learning, with an interest in Reinforcement Learning for social good

Results 4 issues of Ben Glickenhaus

Progressive distillation

6 comment

Starting to work on implementing [PROGRESSIVE DISTILLATION FOR FAST SAMPLING OF DIFFUSION MODELS](https://arxiv.org/pdf/2202.00512.pdf) This is very much a draft PR. So far I've included a toy example in a notebook....

Advantage of ThreadPool over Pool?

I'm just curious why this uses threading instead of multiprocessing in the EvolutionModule? My environment is pretty CPU heavy and my hunch is processes would give more of a speed...

Inline support for Many to Many Objects

1 comment

First off, this is a super cool tool! I'm just starting to explore all the functionality. One thing I noticed is that while it seems to support ManytoMany fields, I...

NaN Loss on train

4 comment

I have been getting NaN losses when I try to train my agent on a game. I tracked it back to the get_batch function in memory - Y (the output...