HalfCauchy should ValueError if _validate_args

[feynmanliang]

Expected: When I run torch.distributions.HalfCauchy(torch.tensor(1.0), validate_args=True).log_prob(-1), I expect a ValueErro because that is the behavior of other distributions (e.g. Beta, Bernoulli).

Actual: No run-time error is thrown, but a -inf log prob is returned.

Fixes #50404

---

This change is 

[facebook-github-bot]

💊 CI failures summary and remediations

As of commit edd20ab (more details on the Dr. CI page):

---

💚 💚 Looks good so far! There are no failures yet. 💚 💚

---

This comment was automatically generated by Dr. CI (expand for details).

Follow this link to opt-out of these comments for your Pull Requests.

Please report bugs/suggestions to the (internal) Dr. CI Users group.

[neerajprad]

Thank you fixing this, @feynmanliang!

[feynmanliang]

Let me fix CI; I think the tests may need a bump

[neerajprad]

Hmm..there is a test failure in test_distributions which is genuine.

Jan 12 01:08:33 ERROR [0.004s]: test_halfcauchy_shape_scalar_params (__main__.TestDistributionShapes)
Jan 12 01:08:33 ----------------------------------------------------------------------
Jan 12 01:08:33 Traceback (most recent call last):
Jan 12 01:08:33   File "distributions/test_distributions.py", line 3234, in test_halfcauchy_shape_scalar_params
Jan 12 01:08:33     self.assertEqual(halfcauchy.log_prob(self.scalar_sample).size(),
Jan 12 01:08:33   File "/Users/distiller/workspace/miniconda3/lib/python3.7/site-packages/torch/distributions/half_cauchy.py", line 54, in log_prob
Jan 12 01:08:33     self._validate_sample(value)
Jan 12 01:08:33   File "/Users/distiller/workspace/miniconda3/lib/python3.7/site-packages/torch/distributions/distribution.py", line 255, in _validate_sample
Jan 12 01:08:33     raise ValueError('The value argument to log_prob must be a Tensor')
Jan 12 01:08:33 ValueError: The value argument to log_prob must be a Tensor

[codecov]

Codecov Report

Merging #50403 (edd20ab) into master (bb97503) will decrease coverage by 0.00%.
The diff coverage is 100.00%.

@@            Coverage Diff             @@
##           master   #50403      +/-   ##
==========================================
- Coverage   80.70%   80.70%   -0.01%     
==========================================
  Files        1905     1905              
  Lines      206755   206757       +2     
==========================================
  Hits       166859   166859              
- Misses      39896    39898       +2     

[facebook-github-bot]

@neerajprad has imported this pull request. If you are a Facebook employee, you can view this diff on Phabricator.

[facebook-github-bot]

@neerajprad merged this pull request in fc9f013.