Running many sims by saving previous sims #1528
Closed
joel-becker
started this conversation in
Squiggle Improvement Proposal
Replies: 2 comments
-
We're working a bit on a Squiggle Server, that might be able to help a bit here. That said, needing 1M sims is kind of expensive, shouldn't be run too often. |
Beta Was this translation helpful? Give feedback.
0 replies
-
We've been making it easier to export samples. We have a simple Alpha API for running Squiggle code, at the moment. We might be able to fairly easily use this with the editor, or similar. Message me again if you need this in the future!. |
Beta Was this translation helpful? Give feedback.
0 replies
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
-
Sometimes, I'm multiplying lots of long-tailed distributions together. I try to cut down a bit to avoid the curse of dimensionality issue (moments of distribution become hard to estimate if you have too many dimensions), but get to a point where I really don't want to cut down any further. I then want to run a large number of simulations to get a stable estimate for the moments of my distribution. At the moment, Squiggle crashes around ~50k sims, which often isn't enough for stable estimates.
One way around this would be to save multiple sets of <50k sims, then combine to find e.g. the mean with 1M sims. This might go hand-in-hand being able to export samples.
Beta Was this translation helpful? Give feedback.
All reactions