Prevent tests from leaking their respective RNG (#4497)

* Add autouse fixture to save and reset RNG in tests

* Add other RNG generators

* delete freeze_rng_state

* Hopefully fix GaussianBlur test

* Alternative fix, probably better

* revert changes to test_models

Co-authored-by: Francisco Massa <fvsmassa@gmail.com>

Author: NicolasHug

test/conftest.py

@@ -1,5 +1,7 @@
 from common_utils import IN_CIRCLE_CI, CIRCLECI_GPU_NO_CUDA_MSG, IN_FBCODE, IN_RE_WORKER, CUDA_NOT_AVAILABLE_MSG
 import torch
+import numpy as np
+import random
 import pytest
 
 
@@ -80,3 +82,26 @@ def pytest_sessionfinish(session, exitstatus):
     # To avoid this, we transform this 5 into a 0 to make testpilot happy.
     if exitstatus == 5:
         session.exitstatus = 0
+
+
+@pytest.fixture(autouse=True)
+def prevent_leaking_rng():
+    # Prevent each test from leaking the rng to all other test when they call
+    # torch.manual_seed() or random.seed() or np.random.seed().
+    # Note: the numpy rngs should never leak anyway, as we never use
+    # np.random.seed() and instead rely on np.random.RandomState instances (see
+    # issue #4247). We still do it for extra precaution.
+
+    torch_rng_state = torch.get_rng_state()
+    builtin_rng_state = random.getstate()
+    nunmpy_rng_state = np.random.get_state()
+    if torch.cuda.is_available():
+        cuda_rng_state = torch.cuda.get_rng_state()
+
+    yield
+
+    torch.set_rng_state(torch_rng_state)
+    random.setstate(builtin_rng_state)
+    np.random.set_state(nunmpy_rng_state)
+    if torch.cuda.is_available():
+        torch.cuda.set_rng_state(cuda_rng_state)

test/test_transforms_tensor.py

@@ -714,6 +714,7 @@ def test_random_apply(device):
 @pytest.mark.parametrize('channels', [1, 3])
 def test_gaussian_blur(device, channels, meth_kwargs):
     tol = 1.0 + 1e-10
+    torch.manual_seed(12)
     _test_class_op(
         T.GaussianBlur, meth_kwargs=meth_kwargs, channels=channels,
         test_exact_match=False, device=device, agg_method="max", tol=tol