update

PaddlePaddle · Oct 8, 2023 · 91f9e01 · 91f9e01
1 parent bffd1ee
commit 91f9e01
Show file tree

Hide file tree

Showing 3 changed files with 15 additions and 23 deletions.
diff --git a/python/paddle/distribution/binomial.py b/python/paddle/distribution/binomial.py
@@ -200,7 +200,8 @@ def entropy(self):
             numpy.ndarray: the entropy for the binomial r.v.
         """
         values = self._enumerate_support()
-        log_prob = paddle.nan_to_num(self.log_prob(values), neginf=0)
+        eps = paddle.finfo(self.probability.dtype).eps
+        log_prob = paddle.nan_to_num(self.log_prob(values), neginf=eps)
         return -(paddle.exp(log_prob) * log_prob).sum(0)
 
     def _enumerate_support(self):

diff --git a/python/paddle/distribution/poisson.py b/python/paddle/distribution/poisson.py
@@ -177,17 +177,8 @@ def entropy(self):
         values = self._enumerate_bounded_support(self.rate).reshape(
             (-1,) + (1,) * len(self.batch_shape)
         )
-        log_prob = paddle.nan_to_num(
-            (
-                -self.rate
-                + values * paddle.log(self.rate)
-                - paddle.lgamma(values + 1)
-            ),
-            neginf=0,
-        )
-        return paddle.nan_to_num(
-            -(paddle.exp(log_prob) * log_prob), posinf=0
-        ).sum(0)
+        log_prob = self.log_prob(values)
+        return -(paddle.exp(log_prob) * log_prob).sum(0)
 
     def _enumerate_bounded_support(self, rate):
         """Generate a bounded approximation of the support. Approximately view Poisson r.v. as a Normal r.v. with mu = rate and sigma = sqrt(rate).
@@ -218,14 +209,14 @@ def log_prob(self, value):
             raise ValueError(
                 'Every element of input parameter `value` should be nonnegative.'
             )
-
+        eps = paddle.finfo(self.rate.dtype).eps
         return paddle.nan_to_num(
             (
                 -self.rate
                 + value * paddle.log(self.rate)
                 - paddle.lgamma(value + 1)
             ),
-            neginf=0,
+            neginf=eps,
         )
 
     def prob(self, value):

diff --git a/test/distribution/test_distribution_binomial_static.py b/test/distribution/test_distribution_binomial_static.py
@@ -31,13 +31,13 @@
     [
         (
             'one-dim',
-            np.array([100]),
+            np.array([20]),
             parameterize.xrand((1,), dtype='float32', min=0, max=1),
         ),
         (
             'multi-dim',
-            np.array([100]),
-            parameterize.xrand((3, 2), dtype='float32', min=0, max=1),
+            np.array([20]),
+            parameterize.xrand((1, 3), dtype='float32', min=0, max=1),
         ),
     ],
 )
@@ -58,7 +58,7 @@ def setUp(self):
             var = dist.variance
             entropy = dist.entropy()
             mini_samples = dist.sample(shape=())
-            large_samples = dist.sample(shape=(5000,))
+            large_samples = dist.sample(shape=(500,))
         fetch_list = [mean, var, entropy, mini_samples, large_samples]
         feed = {
             'probability': self.probability,
@@ -191,17 +191,17 @@ def test_prob(self):
     [
         (
             'one-dim-probability',
-            np.array([75]),
+            np.array([16]),
             parameterize.xrand((1,), dtype='float32', min=0, max=1),
             np.array([75]),
             parameterize.xrand((1,), dtype='float32', min=0, max=1),
         ),
         (
             'multi-dim-probability',
-            np.array([189]),
-            parameterize.xrand((5, 3), dtype='float32', min=0, max=1),
-            np.array([189]),
-            parameterize.xrand((5, 3), dtype='float32', min=0, max=1),
+            np.array([32]),
+            parameterize.xrand((1, 2), dtype='float32', min=0, max=1),
+            np.array([32]),
+            parameterize.xrand((1, 2), dtype='float32', min=0, max=1),
         ),
     ],
 )