Bob Carpenter

icon indicating a website link https://bob-carpenter.github.io icon indicating an email [email protected]

icon company Flatiron Institute icon location New York icon location I work in computational and applied statistics, largely on the Stan probabilistic programming language, autodiff library, and Bayesian inference algorithms.

Results 667 comments of Bob Carpenter

log_prob function on cmdstanr output

Duplicate of #679

Fixed a few typos

@jgabry: Please don't merge this until there is a license statement for what we're doing. The GitHub user agreement means anyone submitting code is agreeing to do it under the...

Extend `get_dims` and `get_param_names` in model_base

> have a quick look at this and evaluate on how difficult this change will be to manage for rstan? On the code side, probably nothing because of the default...

Extend `get_dims` and `get_param_names` in model_base

@bgoodri: Can you avoid passing them? I couldn't tell if you were saying it would be OK for RStan or a problem.

Fix/docs2

Thanks so much for finishing the revised PR. I'm not going to review, because I don't understand the new math lib's C++ well enough. Ideally, someone would have the time...

Fix/docs2

> Are there any objections to committing @rok-cesnovar's grammar suggestions and then merging this? Yes, I think it's better than what we have now and I don't think @SteveBronder has...

opencl question

What would a stan-math-gpu library look like and in what ways would it be better, and who will it be better. Is it a separate translation unit? Does that benefit...

more helpful "PCH file built from a different branch" message

The forum posts I found had instructions for fixing that no longer work. So I put my SEO hat on and formulated a forum post that I hope has the...

Some computations not included in reported total time of output CSVs

@jtimonen. Thanks so much for the careful bug report. For the timing reports, the only solutions that make sense to me are to (a) include all the time taken, or...

Some computations not included in reported total time of output CSVs

I'm totally OK with including total run time and more fine-grained ODE solver reports. I didn't realize until now that they had a lot of extra overhead, but of course...