Caleb Weinreb
Caleb Weinreb
Same here
Hi Scott, thanks for clarifying! I think we landed on a good way of articulating the algorithm over slack. I'll repost here in case others are interested: - Let's assume...
> This looks really neat @calebweinreb! > > One question about the timing results - is that on a cpu or gpu? I remember the behaviour being a bit different...
Hi Adrien, The way I think about the algorithm, each associative scan element is a function $E_{st}: [K] \to [K]$ where $E_{st}(i) \mapsto z_s \sim P(z_s \mid x_{1:t-1}, z_t=i)$. The...
Hi Alex, As far as I know, you are free to upload the files and host them however you want want since the paper is published. Let me know if...
Hi Caleb & Alex, Yes, Caleb is correct. The files are publicly available in multiple formats from the locations linked below. I am not sure if a more specific format/version...