WIP: Added CDF for gamma, inverse gamma and log normal densities #1250
Conversation
numpyro/distributions/continuous.py
Outdated
| @@ -447,6 +451,10 @@ def variance(self): | |||
| def tree_flatten(self): | |||
| return super(TransformedDistribution, self).tree_flatten() | |||
|
|
|||
| def cdf(self, x): | |||
| inv_transform = PowerTransform(-1.0).inv | |||
| return 1 - self.base_dist.cdf(inv_transform(x)) | |||
There was a problem hiding this comment.
nit: I guess you can use cdf(1 / x) for simplicity (and a bit faster)
numpyro/distributions/continuous.py
Outdated
| @@ -845,6 +853,10 @@ def variance(self): | |||
| def tree_flatten(self): | |||
| return super(TransformedDistribution, self).tree_flatten() | |||
|
|
|||
| def cdf(self, x): | |||
| inv_transform = ExpTransform().inv | |||
| return self.base_dist.cdf(inv_transform(x)) | |||
There was a problem hiding this comment.
nit: maybe using cdf(log(x)) for consistency with the above one
|
Hi @omarfsosa, you might want to merge the fixes in #1251 to unblock your PR. |
|
Ah thanks, I was wondering what happened there. I'll merge those. |
omarfsosa
changed the title
|
@omarfsosa There are a couple of failing tests due to numerical issues and a weird failing test for vectorized chain method. But I have addressed them in that PR and it has been merged to master. You might want to merge with |
|
Hi @omarfsosa, could you merge the master branch? CI is not stable recently. We just added a couple of other fixes for CI. I think after merging, this PR should pass CI. |
Just a couple of nice-to-have methods.