Make autogen functions correct for multiple outputs and views (#31990)

Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/31990

This PR does three things:
- Add a new `allow_rebase_history` flag to the differentiable views. If set, trying to rebase their history will raise an error.
- Make sure that the codegen functions verify this flag before doing inplace operations so that they fail before doing the inplace modification.
- Make sure the codegen functions set this flag properly when we don't support rebasing the history of the output.

The codegen change can be found [here](https://github.com/pytorch/pytorch/commit/4bf180caa0e655111b036c8b0419da89fc75671b).

Test Plan: Imported from OSS

Differential Revision: D19409649

Pulled By: albanD

fbshipit-source-id: a2b41c2d231e952ecfe162bdb6bad620ac595703
This commit is contained in:
albanD
2020-01-24 14:29:11 -08:00
committed by Facebook Github Bot
parent 9e59244b53
commit 3ab30753e9
9 changed files with 166 additions and 94 deletions
+2 -2
View File
@@ -105,7 +105,7 @@ TEST(InitTest, CanInitializeTensorThatRequiresGrad) {
ASSERT_THROWS_WITH(
tensor.fill_(1),
"a leaf Variable that requires grad "
"has been used in an in-place operation");
"is being used in an in-place operation");
ASSERT_EQ(torch::nn::init::ones_(tensor).sum().item<int32_t>(), 12);
}
@@ -168,4 +168,4 @@ TEST(InitTest, FanModeLegacyEnum) {
FANMODE_ENUM_LEGACY_WARNING_CHECK(kaiming_uniform_, FanIn, "torch::kFanIn")
FANMODE_ENUM_LEGACY_WARNING_CHECK(kaiming_uniform_, FanOut, "torch::kFanOut")
}
}