Thanks -- you're my hero! Sorry for not seeing that
myself --- I wasn't careful enough in following
the pattern in normal.hpp!
I'll continue working on it and check in the vectorized
version soon. I got sidetracked on writing a new
chapter on clustering models for the manual.
- Bob
On 1/3/13 5:22 PM, Daniel Lee wrote:
> Found it!
>
> In neg_binomial.hpp, on lines 106 and 113, both assignments ("=") have to be increments ("+="). In the vectorized tests,
> the gradients are being overwritten instead of added.
>
>
>
>
> On Thu, Jan 3, 2013 at 5:16 PM, Daniel Lee <
bea...@alum.mit.edu <mailto:
bea...@alum.mit.edu>> wrote:
>
> Ok. I misunderstood the first email. I see the test failures now. I'll let you know what's going on in a few mins.
>
>