OSSForks/pytorch
1
0
Fork 0
mirror of https://github.com/zebrajr/pytorch.git synced 2025-12-06 12:20:52 +01:00
Code Issues Actions 38 Packages Projects Releases Wiki Activity

detect inplace modifications of views earlier (fix #21875) (#46204)

Browse Source 
Summary: Pull Request resolved: https://github.com/pytorch/pytorch/pull/46204

Test Plan: Imported from OSS

Reviewed By: izdeby

Differential Revision: D24259500

Pulled By: bdhirsh

fbshipit-source-id: 223f8a07da4e4121009fc0a8b6760d90eef089b3
This commit is contained in:
Brian Hirsh 2020-10-19 08:56:05 -07:00 committed by Facebook GitHub Bot
parent 0c5cd8c2b9
commit 00c779a92b
6 changed files with 39 additions and 15 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
aten/src/ATen/native/Sorting.cpp
View File

@ -152,7 +152,7 @@ void quantile_impl(
} else if (dim == self.dim() - 1) {
sorted = std::get<0>(self.sort());
} else {
sorted = std::get<0>(self.unsqueeze(-1).transpose_(dim, -1).sort());
sorted = std::get<0>(self.unsqueeze(-1).transpose(dim, -1).sort());
}
// Treat q as a 1D tensor for the following computations

3
test/jit/test_freezing.py
View File

@ -1068,7 +1068,8 @@ class TestFreezing(JitTestCase):
inp = torch.ones(1, 8, 32, 32)
out1 = fmod.forward(inp)
# FIXME: frozen module mutated from outside (original module).
smod.weight[0, 0, 0, 0] += 100.0
with torch.no_grad():
smod.weight[0, 0, 0, 0] += 100.0
out2 = fmod.forward(inp)
out3 = smod(inp)
self.assertNotEqual(out1, out2)

2
test/jit/test_tracer.py
View File

@ -1538,7 +1538,7 @@ class TestTracer(JitTestCase):
x[i, :] = torch.zeros(4)
return x
self.checkTrace(foo, (torch.rand(3, 4),))
self.checkTrace(foo, (torch.rand(3, 4),), inputs_require_grads=False)
def test_trace_checker_inplace_on_view(self):
def foo(x):

29
test/test_autograd.py
View File

@ -3520,6 +3520,16 @@ class TestAutograd(TestCase):
test()
self.assertEqual(dealloc[0], 1)
def test_inplace_view_leaf_errors(self):
# Issue #21875: Fail faster (when we try to modify the view vs. in backward())
x = torch.zeros(1, requires_grad=True)
y = x.view_as(x)
with self.assertRaisesRegex(RuntimeError,
"a view of a leaf Variable that "
"requires grad is being used in "
"an in-place operation."):
y.add_(1)
def test_inplace_view_backward(self):
# Issue #10532: Make sure that this does not raise RuntimeError.
net = nn.Sequential(
@ -3564,13 +3574,10 @@ class TestAutograd(TestCase):
s.backward()
self.assertEqual(s, torch.tensor(1.0))
# Issue 23502: Ensure RuntimeError for modification of SavedVariable.
# Issue #21875: Fail faster (when we try to modify the view vs. in backward())
a = torch.rand(10, requires_grad=True).narrow(0, 0, 10)
b = a.relu_()
c = b.add_(100)
del b
with self.assertRaises(RuntimeError):
c.sum().backward(torch.ones(1, requires_grad=True))
b = a.relu_()
def test_mul_out(self):
a = torch.randn(2, 2, requires_grad=True)
@ -4421,7 +4428,9 @@ for shape in [(1,), ()]:
if fn_id == "view_of_temp":
# This will be fixed after the deprecation cycle and the warning becomes
# an error.
with self.assertRaisesRegex(RuntimeError, "Jacobian mismatch for output 0"):
with self.assertRaisesRegex(RuntimeError,
"a view of a leaf Variable that requires grad "
"is being used in an in-place operation."):
gradcheck(fn, (a, b))
else:
# This works but the custom backward is not called (or called with partial)
@ -4435,7 +4444,13 @@ for shape in [(1,), ()]:
bw_called[0] = 0
ga_nz[0] = True # For the case where the backward is called
with warnings.catch_warnings(record=True) as w:
fn(a, b).backward()
if inplace and output_is_a_view and fn_id != "one_output":
with self.assertRaisesRegex(RuntimeError,
"a view of a leaf Variable that requires grad "
"is being used in an in-place operation."):
fn(a, b).backward()
else:
fn(a, b).backward()
expected_called = 1
expected_ga_nz = True

14
test/test_nn.py
View File

@ -1552,7 +1552,8 @@ class TestNN(NNTestCase):
m = nn.Linear(20, 10).float()
mw = m.weight[:]
m.double()
mw[0][0] = 5
with torch.no_grad():
mw[0][0] = 5
self.assertTrue(mw[0][0].dtype == torch.float)
self.assertTrue(mw._base[0][0].dtype == torch.double)
@ -1565,7 +1566,8 @@ class TestNN(NNTestCase):
m = nn.Linear(20, 10).float()
mw = m.weight[:]
m.double()
mw[0][0] = 5
with torch.no_grad():
mw[0][0] = 5
self.assertTrue(mw[0][0] == mw._base[0][0])
# Test that if `torch.__future__.get_overwrite_module_params_on_conversion() == True`,
@ -3815,8 +3817,9 @@ class TestNN(NNTestCase):
# use sequential to verify nesting
m = nn.Sequential(CustomState())
m[0].param[0] = 10
m[0].sub.weight[0, 0] = 555
with torch.no_grad():
m[0].param[0] = 10
m[0].sub.weight[0, 0] = 555
state_dict = m.state_dict()
self.assertEqual(state_dict["0.serialized"].item(), 11)
self.assertIn("0.sub.weight", state_dict)
@ -11373,7 +11376,8 @@ class TestNNDeviceType(NNTestCase):
m = nn.Linear(20, 10)
mw = m.weight[:]
m.to(device)
mw[0][0] = 5
with torch.no_grad():
mw[0][0] = 5
self.assertTrue(mw[0][0] == mw._base[0][0])
# Test that if `torch.__future__.get_overwrite_module_params_on_conversion() == True`,

4
torch/csrc/autograd/VariableTypeUtils.h
View File

@ -47,6 +47,10 @@ inline void check_inplace(const Tensor& tensor) {
auto diff_view_meta = static_cast<DifferentiableViewMeta*>(impl::get_autograd_meta(var));
// This can throw or warn
handle_view_on_rebase(diff_view_meta);
if (tensor._base().is_leaf()) {
AT_ERROR(
"a view of a leaf Variable that requires grad is being used in an in-place operation.");
}
}
if (var.is_leaf()) {
AT_ERROR(