gaspersic/Gymnasium
1
0
Fork 0
mirror of https://github.com/Farama-Foundation/Gymnasium.git synced 2025-08-23 23:12:46 +00:00
Code Issues Packages Projects Releases Wiki Activity

fix for issue 1256 (Box(low=0, high=255, dtype='uint8').sample() returned zeros) (#1307)

Browse Source
This commit is contained in:
pzhokhov
2019-02-05 17:49:29 -08:00
committed by GitHub
parent 4ceff7dc09
commit 3067a0b890
2 changed files with 47 additions and 29 deletions
Download Patch File Download Diff File Expand all files Collapse all files

3
gym/spaces/box.py
View File

@@ -41,7 +41,8 @@ class Box(Space):
self.np_random.seed(seed)
def sample(self):
return self.np_random.uniform(low=self.low, high=self.high + (0 if self.dtype.kind == 'f' else 1), size=self.low.shape).astype(self.dtype)
high = self.high if self.dtype.kind == 'f' else self.high.astype('int64') + 1
return self.np_random.uniform(low=self.low, high=high, size=self.low.shape).astype(self.dtype)
def contains(self, x):
return x.shape == self.shape and (x >= self.low).all() and (x <= self.high).all()

19
gym/spaces/tests/test_spaces.py
View File

@@ -38,7 +38,7 @@ def test_roundtripping(space):
@pytest.mark.parametrize("space", [
Discrete(3),
Box(low=np.array([-10, 0]),high=np.array([10, 10])),
Box(low=np.array([-10, 0]), high=np.array([10, 10]), dtype=np.float32),
Tuple([Discrete(5), Discrete(10)]),
Tuple([Discrete(5), Box(low=np.array([0, 0]), high=np.array([1, 5]), dtype=np.float32)]),
Tuple((Discrete(5), Discrete(2), Discrete(2))),
@@ -66,3 +66,20 @@ def test_equality(space):
def test_inequality(spaces):
space1, space2 = spaces
assert space1 != space2, "Expected {} != {}".format(space1, space2)
@pytest.mark.parametrize("space", [
Discrete(5),
Box(low=0, high=255, shape=(2,), dtype='uint8'),
])
def test_sample(space):
space.seed(0)
n_trials = 100
samples = np.array([space.sample() for _ in range(n_trials)])
if isinstance(space, Box):
expected_mean = (space.high + space.low) / 2
elif isinstance(space, Discrete):
expected_mean = space.n / 2
else:
raise NotImplementedError
np.testing.assert_allclose(expected_mean, samples.mean(), atol=3.0 * samples.std())